IT teams are growing increasingly concerned about AI-related risks amidst continued adoption, according to new research.

Findings from Heimdal’s State of AI Risk Management in 2026 report show AI tools are now commonplace across most IT estates, with teams often running several at one.

ChatGPT, for example, runs in nearly three-quarters (71%) of UK IT environments, while Microsoft Copilot is present in 68%. These same figures are reflected across the Atlantic, with US-based IT teams often using a combination of multiple different AI solutions.

AI tools are playing a key role in reducing workloads, the study noted, which IT teams highlighted as the most positive benefits of the technology.

Nearly three-quarters of IT and security teams said they lose around a quarter of their week to “repetitive, low-value work” - and AI is helping reduce that manual toil.

Indeed, teams facing the highest levels of operational load are often ranked among the most optimistic when it comes to AI. More than half (59%) of US teams said they expect AI to alleviate pressure, while 55% in the UK expect the same.

AI-related risks are haunting IT teams

Despite this optimism, a key recurring concern among IT leaders is that controls and security capabilities haven’t kept pace with the rate of adoption. Heimdal noted that only four-in-ten teams rate their security stack as “ready for AI-related risk”.

Teams are increasingly concerned about data leakage, for example, with 56% of UK respondents highlighting this risk. Visibility is a major issue in this regard, Heimdal found.

UK teams with full visibility into AI use were most likely to flag data leakage as a leading concern, compared to just 27% of those with no visibility. In the US, 59% of teams with full visibility also highlighted data leakage as a key concern.

While concerns over potential risks typically mount as AI tools are integrated, Heimdal noted that unauthorized AI use, or ‘shadow AI’, is also a recurring problem for enterprises.

Heimdal specifically highlighted the Salesloft Drift breach in August 2025 as a key example of how poor AI-related visibility can impact organisations.

The incident saw threat actors steal OAuth tokens for Drift’s AI chatbot integration with Salesforce, using these to extract data from several hundred Salesforce instances.

A host of organisations, including Cloudflare, Palo Alto Networks, and Zscaler were impacted in the attacks.

“Drift was the AI tool. Salesforce held the data,” the company noted in a blog post.

“Most of the affected teams had never personally provisioned Drift,” it added. “A third-party AI chatbot, plugged in through an OAuth grant few had recently reviewed, became the way in.”

Contrasting priorities

Perception of AI-related risk among frontline practitioners and executives is a major problem, according to Heimdal. Indeed, “executive confidence” in AI security is a repeated point of friction when it comes to governance and risk management.

In the US, for example, 29% of executives said AI risk is under control, yet just 7% of practitioners agreed. In the UK, meanwhile, these figures stand at 18% against 11%.

Adam Pilton, cybersecurity advisor at Heimdal, said this shows many organizations still aren’t fully aligned on how they manage AI risk.

"Misplaced confidence is one of the most dangerous things in security. This data shows executives are far more confident that AI risk is under control than the evidence supports. Most of the conversation right now is about productivity, when the bigger question is how AI can be turned against the business,” he said.

“The report shows the gap between how secure leaders feel and how secure they actually are.”

FOLLOW US ON SOCIAL MEDIA

More than three-quarters of utilities organizations were hit by cyber attacks involving outdated software or unavailable patches on legacy equipment over the last year.

At 77%, it was the most common type of cyber incident facing the sector, according to Bridewell's Cyber Security in Critical National Infrastructure Report 2026.

And the most common effect was IT disruption or outages, affecting 47% of organizations, despite the fact that 99% of respondents described themselves as resilient after their worst cyber attack.

A further 42% said incidents had resulted in increased cybersecurity spending, while 35% experienced data loss, 34% reported revenue loss, and 32% suffered disruption to production or services.

Supply chain attacks take the longest to respond to, at 9.9 hours on average, followed by data theft or disclosure at 8.4 hours and unauthorised access at 7.6 hours.

The utilities sector is particularly hampered by the need to secure ageing operational technology and infrastructure that weren't designed to withstand modern cyber threats, as critical assets can't be updated or taken offline as easily as traditional IT environments.

"Many of the systems underpinning essential utilities services were designed to operate for decades in environments that were never intended to be connected to modern digital networks," said Sam Thornton, COO at Bridewell.

Beyond legacy infrastructure, phishing and business email compromise remain widespread, affecting 76% of utilities organizations in the past year. Malware affected almost as many, at 74%, while more than seven-in-ten experienced unauthorized system access.

The main concern for utilities organizations is data protection and privacy, cited by 46% of survey respondents. Managing AI-related cyber risk and the ability to quickly detect incidents were close behind, reflecting growing concerns around emerging technologies and increasingly sophisticated attacks.

Utilities organizations are also unconfident when it comes to data breach notification requirements, cited by 42%, cybersecurity measures for data protection at 39%, and third-party due diligence at 38%.

And regulation is now the primary driver of cyber security maturity within the utilities sector, cited by 36% of respondents - ahead of both the evolving threat landscape and customer demand for improved security, and highlighting the growing influence of frameworks and compliance obligations on cyber security investment and decision-making.

"As utilities providers continue to modernize and connect operational systems, managing the gap between legacy infrastructure and modern security requirements is becoming one of the sector's biggest cybersecurity challenges," said Thornton.

Bridewell recommends that utilities organizations improve the visibility of assets across both IT and operational technology environments to identify unmanaged or vulnerable systems.

They should prioritize patch management and vulnerability remediation based on operational risk and criticality, conduct regular incident response exercises to ensure teams can respond effectively during a live cyber incident and strengthen monitoring and detection capabilities to reduce the time taken to identify and contain threats.

They should also review third-party and supply chain security arrangements to ensure critical partners meet appropriate security standards.

"In the utilities sector, the consequences of a cyber attack extend far beyond IT," said Thornton. "When critical systems are disrupted, the impact can be felt by customers, communities and the wider economy, making cyber resilience a business-critical priority."

Tens of thousands of Fortinet devices may have leaked credentials after a sophisticated automated brute force campaign that has security researchers calling for affected companies to reset their passwords now.

Threat intelligence company Hudson Rock posted a blog detailing what it's named Fortibleed, saying 73,932 firewall devices had seen credentials leaked, impacting a long list of organisations – if yours is on the list, change your passwords now.

In a statement sent to ITPRO, Fortinet disputed the details, saying this wasn't the result of a fresh hack and that anyone following best practices was already safe. "Fortinet is aware of a reported third-party credential-harvesting campaign targeting Fortinet firewalls and VPN gateways. We are committed to safeguarding our customers, and we diligently and continuously monitor threat actor darknet activity. Based on our initial analysis, the data involved is likely a resharing of data from previous incidents, as well as brute forcing of credentials, and not related to any current incident or advisory."

Hudson Rock argued the incident isn't that simple – and claimed it's impacting a huge number of companies.

"The group's methodology goes beyond simple credential reuse," the blog post notes. "They actively intercept SSL VPN authentication hashes and crack them using a massive, dedicated 45-GPU cluster managed via Hashtopolis. Once the perimeter is breached, the operators systematically pivot directly into internal Active Directory environments to establish deep network persistence."

The techniques aside, it's the size of the incident that's worrying, Hudson Rock said. "The scale of this breach touches nearly every sector of the global economy, sparing no industry," the blog post added. "The threat actors have built a verified database of working credentials for some of the largest enterprises on the planet."

What's happening?

The reporting on this incident started with a post by security researcher Volodymyr Diachenko, who spotted the mass exploitation in action last week, noting tens of thousands of companies' names were listed.

"Crooks use sophisticated hashcracking approach to get then plaintext passwords from the Fortigate configs and use them consequently in the internal network movement and takeover," he said at the time.

Another security researcher, Kevin Beaumont, then examined the incident. He said the data appeared to be legitimate and that 75,000 devices were impacted, with most still online and the majority from Fortinet. Beaumont said he had worked with some of the companies and could confirm the data.

"The data comprises of roughly 50% of all Fortinet firewall devices facing the internet, based on polling from Shodan," he noted.

The data could allow attackers to log in remotely and access the firewall and network it was protecting, Beaumont said. "They can also change settings, including security controls, and make backdoor users," he wrote.

What can companies do?

Fortinet said its customers following best practices and recent advice should be safe. "Organizations that follow routine best practices, including regularly rotating security credentials and enabling multi-factor authentication, as per guidance in this March blog, face minimal risk from credential compromise detail referenced in the reporting."

Beaumont advised companies to check if they were affected via the Hudson Rock list, and if so, immediately rotate all admin credentials, looking at prior logins for suspicious activity. But he said to "assume compromise" as it's unclear how long the data seen has been in circulation.

If found to be compromised, the entire device may need to be replaced as settings may have been altered allowing for a backdoor to be installed, Beaumont added. Devices should be upgraded to the latest FortiOS version, and admins should login to change passwords. Generally, he advised the FortiOS management interface to not be exposed to the internet unless strictly necessary, and to implement multi-factor authentication.

Hudson Rock echoed that advice, adding that companies should monitor for stolen credentials to spot them before they're weaponised against your network.

Hudson Rock added that one "alarming detail" from the breach was how many complex passwords were successfully compromised, noting that IT departments lean on rigid password rules in a bid for protection. "However, complexity is completely neutralized when passwords are recovered in plaintext," the company noted.

Hudson Rock added: "This massive incident serves as a glaring reminder that exposed network gateways combined with reused or stolen credentials are an attacker's dream."

The overwhelming majority of cyber attacks on critical infrastructure are coming from hostile states, the UK's cyber chief has warned.

Speaking at the Royal United Services Institute's (RUSI) Annual Security Lecture, Richard Horne, CEO of the National Cyber Security Centre (NCSC), said the organization had handled more than 200 cyber incidents affecting the UK's critical national infrastructure and its supporting ecosystem over the last year. Around 75% were believed to be linked to state actors, particularly Russia, China, and Iran.

"We know that adversaries are prepositioning today, establishing footholds within technology that underpins critical national infrastructure that could enable rapid exploitation, to cause mass disruption in a time of conflict," he said.

"The highest profile example of this was a campaign often referred to as Volt Typhoon against largely US critical national infrastructure, which was attributed in 2024. And we are seeing our critical infrastructure being targeted, regularly finding and stopping breaches, before their intent becomes clear."

Horne broke the threat down into 'near', 'mid,' and 'far' spaces, with the far space representing the adversary's home turf, systems, tooling, and networks. Here, he said, the UK and its allies bring pressure to bear through intelligence collection, sanctions, law enforcement action , and offensive cyber operations to disrupt and degrade their capability at source.

In the mid space, efforts are concentrated on hardening cloud, technology, and telecommunications infrastructure, and by disrupting adversary positions within those environments.

"The reality is much of this space is in private hands," he said. "Which means success here demands genuine collaboration between government and private sector, which is at the heart of our approach in the NCSC."

But, he said, it's the near space – the defense and resilience of the organizations and systems being targeted – where most action is probably required. The rise of AI is an important factor here, he said.

"Recent developments of frontier AI models have demonstrated their effectiveness at finding inherent vulnerabilities in the technology we rely on," he said.

"Our latest assessment shows that by 2028, it is highly likely that AI-Cyber capabilities will be used by attackers against known vulnerabilities in legacy technology in our critical national infrastructure."

British organizations should take note, said James Neilson, SVP of global at OPSWAT.

"The daily scale of hostile activity against the UK is vast, and until the NCSC revealed those figures, the threat and danger facing critical infrastructure was far greater than most businesses realized," he said.

"Many organizations neglect to secure data that moves in and out of their OT networks. By controlling data flows and scanning files in transit, organizations can detect and neutralise hidden malicious payloads before they infiltrate critical systems."

Horne called on organizations to strengthen cyber resilience by focusing on three core capabilities: understanding their exposure to threats, building stronger defences based on proven security fundamentals, and ensuring they can continue operating and recover quickly after an attack.

"By making our environment harder for adversaries to operate in, and engaging in the contest better, we can play an important part in altering potential adversaries' options and deterring conflict," he said.

The National Cyber Security Centre (NCSC) recently declared passwords to be fundamentally flawed, so what are the concerns and what can be done about them?

Access and identity management are an essential part of our online lives. They protect our data and ensure sensitive information is only shared with the appropriate people. However, following recent password leaks, billions of access credentials have been exposed, thereby fundamentally weakening the systems relying on them.

Passwords have been used to confirm identity for thousands of years, for example, by guards and sentries to identify friends and official visitors. However, even then, passwords were prone to interception and being used by enemies to falsify identity.

The problem with passwords

The key problem with passwords is that they do not confirm identity, only that someone knows the correct response. The recent leaks have been compounded by the poor password hygiene followed by many users, who use the same password credentials for multiple platforms and services.

Furthermore, recent advances in high-performance computing (HPC) and quantum computing have meant that computers are becoming increasingly powerful and able to crack passwords in a far shorter time than was previously possible.

In 2024, China announced it was able to decrypt 50-bit RSA encryption using quantum computing. Although modern encryption is a minimum of 2048 bits, the research is a fascinating proof of concept, showing where the technology is heading and the implications for the future of cybersecurity.

As the processing capabilities of modern computers continue to grow rapidly, the recommended minimum length and complexity of passwords are becoming longer. It is now recommended that passwords be at least ten characters long, with a mix of letters, numbers, and symbols, and not be a name or word from a dictionary.

Compounding the issue is that not everyone follows appropriate password hygiene, such as not using the same password credentials across multiple accounts or avoiding words/phrases that have a personal connection. Furthermore, the most common passwords are still “123456”, “admin” and “12345678”. At this point, we may as well just give the bad actors our keys.

“The VIPs are the worst security users in the company – they don't want to type even eight characters. I saw in the past some CEOs who are asking their IT people to have only three characters as a password,” says Jean-François Aliotti, co-founder of Almond.

“Now, because of all the leaks that we have seen, there are passwords leaked everywhere. Some of my passwords have been leaked. I use a unique password for each access I have. It's a rule that I follow strictly, and I use password managers for that. But most people don't do that – they have an Excel file with all their passwords, or they have the same password everywhere.”

Regular changes of passwords are commonly enforced, especially for business account login details. The recommended duration varies depending on the sensitivity of the data, but a password change every three to six months is the most common requirement. However, the NCSC has argued against changing passwords due to the potential vulnerabilities it causes, as users can be tempted to rely on passwords that are easier to remember.

It is no longer recommended that access security and identity management be solely reliant upon a password. Instead, there needs to be a layered approach to security, with multiple levels of authentication before granting access.

Alternative solutions

The most common form of additional confirmation is Multi-Factor Authentication (MFA), whereby short-term single-use codes are sent to personal devices held by the user. Codes are typically sent via text, email, or authenticator app, but could also be generated by a 2FA token. However, if someone has already gained access to the secondary device or token, they will be able to confirm the additional verification.

Emails are potentially the most vulnerable form of MFA, as they are equally reliant on passwords, and many personal email accounts are not as strongly protected as they could be.

Biometrics (fingerprints, facial recognition, and voice recognition) are unique to each person, but are not as strong as many believe. Fingerprints can be forged, and voice recognition can be fooled using high-definition recording. Facial recognition can easily be bypassed if a user is caught off guard, as anyone with teenagers will know when friends ‘borrow’ their phones.

“Biometrics are a good thing, but not alone. If someone stole your fingerprint, it's over. You can change your password, but you cannot change your fingerprints,” says Aliotti.

“Biometrics alone are quite dangerous, because if they are stolen, then it's over.”

Passkeys instead of passwords

An alternative authentication system is passkeys. Although the technology is comparatively new, the NCSC has recommended that people use passkeys instead of passwords.

“Adopting passkeys wherever you can is a strong step towards a safer, simpler login experience, and I am pleased that we can now support uptake,” according to Jonathon Ellison, Director for National Resilience, NCSC.

“The headaches that remembering passwords have caused us for decades no longer need to be a part of logging in where users migrate to passkeys – they are a user-friendly alternative which provide stronger overall resilience.

“As we aim to accelerate the UK’s cyber defences at scale, moving to passkeys is something all of us can do to improve the security of everyday digital services and be prepared for modern and future cyber threats.”

When a user seeks to confirm their identity, a push notification is sent to their smartphone. Once their device, such as a smartphone, has been unlocked, a unique passkey is created and sent to the platform/website/service they wish to access, confirming their identity.

Unlike MFA, which relies on traditional methods of user verification, this method does not rely on login information or biometric data being transmitted, thus mitigating interception and key-logging attacks.

“MFAs will continue to be deployed, but what we are seeing right now is that passkeys are the best way, but it will take a lot of time to deploy them at a large scale right now,” says Aliotti.

“Pass keys will be more and more used, and we hope that it will be the dominant way for credentials, as we don't have any other system right now that we are seeing as a brand-new thing.”

The decreasing effectiveness of passwords means they are no longer viable as a sole form of access management. Instead, a layered authentication process, where users need to prove their identity through two or more methods, is strongly recommended.

Furthermore, given the inherent weakness of passwords overall, the robust nature of passkey technology means that passkeys are the NCSC’s recommended access management protocol.

Goldilock Secure has signed a new distribution partnership with Frame Communications, as the cybersecurity vendor aims to expand the availability of its FireBreak technology across Ireland.

The agreement will see Frame leverage its network of resellers, MSPs, and telecommunications partners to distribute the FireBreak platform to organizations operating across enterprise, telecoms, and critical infrastructure environments.

According to Goldilock, the partnership comes amid growing concern over increasingly automated and AI-driven cyberattacks that are accelerating the discovery and exploitation of software vulnerabilities.

The firm's FireBreak solution is designed to reduce this cyber risk through hardware-enforced network disconnection capabilities that physically isolate systems when threats are detected.

The appliance-based platform allows businesses to sever connectivity at layer 1, isolating compromised network segments and helping prevent attackers from reaching critical infrastructure or backup environments.

Steven Brodie, Goldilock's chief revenue officer and head of partnerships, described the new partnership with Frame as a "significant step" in the firm's expansion strategy.

"Frame is a strong addition to our growing global partner network, which has expanded by more than 40 channel partners in just the past six months," he commented. "Their extensive and well-established presence in Ireland, combined with strong relationships across the market, will play a key role in accelerating the adoption of FireBreak in the region."

To support the FireBreak rollout, the companies will work together to provide product enablement, solution design, deployment assistance, and technical support across the partner ecosystem.

The aim is to provide channel partners and customers with a focused, end-to-end pathway to position, sell, deploy, and maintain the solution.

Looking ahead, Goldilock said partners and customers can expect to see increased engagement across Ireland, including increased availability of FireBreak through local resellers as well as targeted opportunities within the telecom sector.

"As modern cyberattacks grow more complex and sophisticated, our customers and partners are looking for practical ways to reduce cyber risk, as well as the ability to isolate and contain threats without taking their entire network offline and impacting operations," commented Frame managing director Gavin McGowan.

"FireBreak is a perfect fit, and we look forward to working together to bring this solution to organisations across Ireland."

Research shows that a third of organizations experienced a data sovereignty incident last year. It is not a case of blind ignorance, though. Indeed, our own Data Sovereignty Report found that 44% of respondents describe themselves as “very well informed” about data sovereignty requirements.

Businesses know the rules. Yet, one in three of them got hit by a sovereignty incident anyway. That gap is the single biggest commercial opportunity in the UK channel right now. It’s getting worse, and the businesses that need help the most are the ones least equipped.

Why does this matter?

Customers don’t buy sovereignty from a vendor slide deck. They buy it from a trusted partner that maps their data flows, identifies where the architecture can’t enforce the policy promises, and builds a remediation plan that passes the audit.

Over four-in-ten (44%) businesses flag concerns about whether their cloud providers can genuinely guarantee sovereignty. Those concerns are well-founded, but the question most customers are asking is the wrong one. It shouldn’t be “where is my data stored?

Ring-fencing data by geography is neither new nor technically difficult. What is far harder, and most customers have never genuinely confronted, is the question of legal jurisdiction.

Consider the architecture that many on these shores believe is sovereign. A major US cloud provider may operate a German-based subsidiary, staffed by EU nationals, marketed explicitly as a sovereign offering. But the parent company remains subject to US law, and no subsidiary structure changes that. A lawful US warrant, a trade embargo, or an executive order doesn’t stop at the border of a local data centre.

Plus, events that would have seemed far-fetched a few years ago (sweeping trade disputes, unilateral policy shifts, foreign data access demands) are no longer theoretical. They are the operating environment of today. And if any of those scenarios materialise, clients and MSPs relying on a geo-residency promise could face real, material exposure.

Target the mid-market

It is the mid-market where the real urgency lives. Sovereignty maturity generally scales with organization size. Among companies with over 20,000 employees, roughly 45% spend above £5 million annually. At the other end, organizations with 500 to 999 employees sit at just 19% in high-tier spending.

Large enterprises often have internal sovereignty teams and dedicated compliance architects. Mid-market organizations, however, have the same regulatory obligations and enforcement exposure, yet only a fraction of the resources. They are the ones that need a partner who can deliver sovereign infrastructure without requiring them to hire a team of specialists to run it.

And time is ticking. GDPR fines now exceed €5.6 billion, and the EU AI Act introduces penalties up to €35 million or 7% of worldwide turnover. For a UK business operating in Europe post-Brexit, the regulatory surface area has never been larger.

Four questions to consider...

These are the key questions you need to get your customers asking themselves.

• Which legal jurisdiction ultimately governs our data? A cloud provider can locate a data centre here and market it as a sovereign offering. But they are still subject to the laws of the country where the parent company is headquartered. If a lawful warrant, a trade dispute, or a government access demand lands on that parent company, the local subsidiary’s address offers limited protection. Jurisdiction follows the entity, not the building.
• Who controls the encryption keys? If the provider retains the ability to decrypt customer data, the customer doesn’t have sovereignty. They have a residency promise with a legal back door. Sole encryption key ownership, retained within the customer’s environment, is the line between sovereignty that holds and sovereignty that folds under a government access request.
• Where is data processed, not just stored? Cloud platforms can store data here in the UK, yet process it abroad without the customer knowing. For regulated industries, that invisible border crossing is a compliance violation waiting to happen.
• Can you prove it? Regulators and procurement teams no longer accept “we believe we’re compliant.” They want immutable audit trails, residency logs, and compliance documentation produced on demand. That’s the shift from stated compliance to provable control.

Channel partners should look at this as an architecture engagement. Map the data flows. Deploy a platform that enforces residency at the infrastructure level, retains key custody in-jurisdiction, and generates audit evidence. That’s a services-rich, high-value, recurring-revenue conversation. Plus, it renews, because sovereignty isn’t a project. It’s a permanent operating condition.

The conversation to have

Partners winning the sovereignty conversation are the ones leading with the jurisdiction question, targeting the mid-market, and building sovereignty practices that go beyond the data map.

The data doesn’t lie. What separates the firms that avoided incidents from those that did is operational depth. Architecture, controls, and evidence.

However, what will separate the channel partners that will win in the future will be something more foundational. It’s all about the willingness to have the conversations that the vendors won’t.

Data belonging to around 450,000 present and former Nottingham University students has been compromised in an attack claimed by the ShinyHunters ransomware group.

The data is believed to include contact information, including names, email and postal addresses, course information, student and staff IDs, financial information, and national insurance numbers.

The university said that when the incident was detected, it immediately took the affected systems offline.

"We are working to understand the data that has been accessed and have contacted those students and alumni affected directly. We are working closely with Action Fraud, the Information Commissioner’s Office, and other regulatory bodies," it said in a statement.

"We will remain in contact with those directly impacted and will continue to provide updates as the situation develops."

ShinyHunters claims responsibility

The attack has been claimed by the ShinyHunters gang, which said on its dark web leak site that it had accessed more than 40GB of data relating to students at the university's Malaysia and China campuses, as well as the main Nottingham campus itself.

The group said it wasn't bluffing and that 'the inevitable' would happen if a ransom wasn't paid – something that the government prohibits universities from doing.

This is the latest in a string of attacks against the higher education sector, which is often seen as a soft target.

"One thing we can say for certain is that the higher education sector is at increased risk of attack today and it is vital these organizations take steps to improve their defenses. Universities are highly valuable to an array of attackers, some looking to steal IP or research, or others looking to monetise on data," said Keven Knight, CEO of Talion.

"The sector needs to recognise these risks and take urgent action to improve the defences of their environments, either through their own internal resources, or by partnering with security experts that specialise in the sector."

University criticized for response

The university has come in for criticism over the way it has handled the breach, particularly with regard to the timeframe.

"Most concerning is the claim that attackers remained undetected in the University of Nottingham’s systems for over a week, giving them ample opportunity to access additional data or move laterally through the network," said Adam Boynton, senior enterprise strategy manager at Jamf.

Brian Higgins, security specialist at Comparitech, suggested the university failed to give students as much information as it should.

"Apart from reporting it to the Information Commissioner’s Office, who can’t actually do anything about it, and reassuring themselves that they take security seriously, at this stage there is more information coming from Have I Been Pwned and ShinyHunters themselves about what’s going on," he said.

"Users, customers and learners deserve better from those entrusted with their data in our current digital society. A few press statement platitudes are useless when all of your information is for sale on the dark web because somebody else leaked it there."

FOLLOW US ON SOCIAL MEDIA

Supply chain cyber risk is now a top concern for managed service providers (MSPs) and their customers, according to new research from CyberSmart.

The cyber risk management provider’s 2026 MSP Survey found that 43% of MSPs and their customers experienced a cyber incident caused by or originating from a supplier or third-party vendor during the last 12 months.

Conducted by OnePoll, the research gathered responses from 350 MSP leaders across the UK and Ireland spanning a range of industries and customer sizes.

The results suggest that MSPs are becoming increasingly exposed to supply chain risk due to their privileged access to customer environments – making them attractive targets for cyber criminals seeking to use them as a gateway to potentially hundreds of organizations.

Among the respondents that experienced a supply chain incident, 39% said the breach affected only the customer, while 16% said it only affected the MSP. Meanwhile, 39% said both were impacted.

According to CyberSmart, this means over half (55%) of incidents involved the MSP in some capacity.

Despite this figure, the research found that 55% of MSPs still do not monitor for supply chain risk – while over a third (37%) only assess risk quarterly and 11% do so annually.

In terms of challenges, MSPs cited managing and enforcing security requirements in contracts (39%), third-party risk assessment and monitoring (37%), and the cost of securing and monitoring supply chains (36%) as the biggest hurdles.

“Supply chain risk has become a central concern for MSPs and SMEs as cybercriminals increasingly target interconnected business ecosystems,” commented CyberSmart CEO and co-founder Jamie Akhtar.

“MSPs sit at the centre of these environments, which means a single weak link can have far-reaching consequences for customers, suppliers and partners.”

Growing regulatory pressure

Elsewhere, the report also explored MSP preparedness ahead of the UK’s Cyber Security and Resilience Bill (CSRB), which was introduced back in November 2025 and brings providers into the scope of formal regulation for the first time.

The CSRB includes mandatory security requirements, stricter incident reporting, and greater accountability as MSPs become increasingly critical components of national cyber resilience.

According to the findings, 96% of respondents said they felt prepared for the legislation to a certain extent, while 45% described themselves as fully prepared.

Addressing key CSRB concerns

However, MSPs pointed to operational and organizational concerns rather than technology and software limitations as the biggest barrier to readiness.

Instead, they cited skills (41%), clearer customer expectations (41%), stronger support for managing third-party risk (41%), as well as better-defined roles and liability (39%) as key requirements going forward.

Increased liability and legal exposure emerged as the biggest concern linked to the new legislation, noted by 42% of the survey’s participants, with MSP leaders expressing concern over undefined accountability and how risk will be operationalized in practice.

Improving accountability and resilience

Despite these concerns, 77% said they believe CSRB goes far enough in helping to protect supply chain organizations – including MSPs themselves – from cyber risk.

When it comes to what can be done to improve protection for MSPs, participants highlighted clearer guidance and best practice standards (54%), stronger protections around shared liability (52%), and clearer regulatory frameworks specifically for MSPs (51%) as the top three potential improvements.

“What our research shows is that the industry understands the need for greater accountability and resilience, but MSPs also need clearer guidance, shared responsibility and continuous risk visibility to make that possible in practice,” Akhtar added.

FOLLOW US ON SOCIAL MEDIA

The JDY botnet is back and expanding via attacks on unpatched routers, cameras and other edge devices.

According to a report from Lumen's Black Lotus Labs the JDY botnet now makes up 1,500 compromised small office and home office (SOHO) devices, as well as edge and Internet of Things (IoT) devices, and is used by Chinese state-backed hackers including Volt Typhoon as a scanner to spot exposed services for exploitation.

JDY was first spotted back in 2023 as part of an investigation into the KV botnet, which was used for covert data transfer while JDY was focused on scanning and reconnaissance. After KV was taken down last year by the US government, JDY remained an active threat, Black Lotus Labs noted in a report.

Now, the JDY botnet has doubled in size, with compromised devices located across Europe and Asia, though the majority are in the US. The network is used to scan a range of targets for weaknesses, though the US military is a clear focus.

"The expansion of the JDY botnet underscores how China‑nexus threat actors are scaling reconnaissance as a core enabler of exploitation," Black Lotus Labs said in a blog post.

"By distributing scanning and fingerprinting across thousands of compromised SOHO and IoT devices, operators can rapidly identify vulnerable infrastructure and targets of interest while evading traditional, IP‑based defenses."

Gabrielle Hempel, Security Operations Strategist at Exabeam, noted that discussions around botnets normally focus on data theft, but there's more to consider.

"We spend a lot of time talking about nation-state actors stealing information, but the scarier reality is that many of these operations are designed to establish positioning and persistence," she said.

"If geopolitical tensions ever escalate, having access already in place is far more valuable than trying to gain it during a crisis. Persistent access provides intelligence collection opportunities today and potential disruption options tomorrow."

Targeting unpatched edge devices

Previously, JDY focused on two Cisco router models but has now expanded its botnet to include a range of manufacturers. Using edge devices helps the botnet's activity blend into regular traffic, the security lab added.

Devices aren't added to the JDY botnet at random, however. Indeed, the attackers are looking for specific models with known exploitable flaws.

"Black Lotus Labs found that JDY botnet operators target specific devices for scanning and reconnaissance, rather than conducting widespread, indiscriminate scanning," the post said. "Most notably, there was a selective increase in scans of Fortinet equipment immediately after the disclosure of a new vulnerability, indicating the ability and intent to find and exploit vulnerable devices before patches are widely applied."

Hempel noted that JDY continues the trend of attackers focusing on easy to exploit edge devices that are often missed in security efforts.

"As we have seen with many recent attacks, campaigns like JDY don’t rely on the sophisticated zero days that everyone loves to talk about, but leverage poorly maintained edge devices, exposed infrastructure, and slow patching," Hempel added. "It’s the low-hanging fruit that they are after to get in.

What should enterprises do?

Given that, it's no surprise that Black Lotus Labs advises companies to follow existing best practice for routers, firewalls, and IOT devices: install patches for known flaws, run security updates, and regularly reboot.

Beyond that, the security lab advised enterprises to adopt the Secure Access Service Edge (SASE) architecture or similar solutions to reduce the attack surface and implement existing guidance from national security bodies about how to mitigate against Volt Typhoon and China-Linked threat groups.

"The JDY botnet underscores the risk of relying on traditional IP-based security controls such as geofencing, IP reputation-based detection, and static blocklists," the security lab added.

"The large number of US-based SOHO and IoT devices that comprise the botnet allows operators to blend in with legitimate user traffic, making malicious scanning and reconnaissance activity harder to detect."

FOLLOW US ON SOCIAL MEDIA

Cyber criminals are exploiting AI hype to impersonate the branding of AI platforms such as ChatGPT, Microsoft Copilot, DeepSeek, and Anthropic’s Claude, according to new research.

Microsoft Threat Intelligence said it's observed an uptick in phishing, malvertising, and search engine optimization (SEO)-driven attacks that ultimately lead to credential theft, financial fraud, or malware infection.

Campaigns focus on highly anticipated launches or emerging trends, using tried-and-tested tactics such as urgency-driven messaging, abuse of trusted services, and multi-stage redirection chains that require user interaction to evade detection.

"While traditional lures like invoices, payment notifications, or delivery alerts remain effective and continue to be widely used, AI-themed lures reflect a shift in social engineering that is likely to persist as a long-term tactic used by threat actors, from cyber criminal groups to nation states," the company warned.

ChatGPT users in the crosshairs

In one example, Microsoft said it had observed a ChatGPT-themed phishing attack delivering malicious URLs which led to phishing pages that collected credit card and personal information such as names and addresses.

The emails used the sender display name ChatGPT and the subject line: “To ensure your ChatGPT Plus continues to work – please update your payment method”.

This phishing activity, which consisted of 4,500 emails sent to targets in South Africa, was part of a broader campaign using similar themes and infrastructure that delivered as many as 100,000 emails on a single day to targets in Switzerland, Austria, and South Africa.

Microsoft noted the campaign affected a broad range of industries, including higher education and professional services.

Thousands targeted in a Claude-themed phishing attack

In another example, security experts spotted a phishing campaign impersonating Anthropic-branded services to target users with account-related lures tied to the Claude AI platform.

The campaign sent phishing emails to targets across more than 2,000 organizations, mainly in the US, UK, and India.

"The campaign used enforcement-themed messaging claiming that the recipient’s account was in violation of acceptable use policies and required immediate action," the company noted.

"The emails impersonated Anthropic’s popular AI service Claude using the display names Anthropic Teams and Anthropic PBC, masquerading as legitimate account-related communications. Subject lines followed a consistent structure of 'Claude Appeal Request' combined with date elements."

DeepSeek malvertising is a growing threat

Other examples included malvertising campaigns that use AI-themed terms such as 'Awesome AI Windows Plugin' and 'Flux Pro AI' in social engineering lures, and fake DeepSeek V4 installers on GitHub that delivered Vidar Stealer.

"Within hours of DeepSeek previewing their latest version, V4, attackers created a fake GitHub organization and repository. They copied real branding and benchmark data, added AI and SEO-search-friendly content, and pushed malicious archives that looked like installers," explained John Bruggeman, vCISO at CBTS.

"What the attacker did was not particularly exotic, but it was well timed and convincingly packaged. A user searching for the newest model could very easily end up in the wrong place, especially because the malicious repository showed up in GitHub, Google, Bing, or AI-assisted search results. The search results added legitimacy to the malware."

Remain vigilant

To counter these rising threats, Microsoft advised customers to configure automatic attack disruption in Microsoft Defender XDR, enforce multi-factor authentication (MFA) on all accounts, use the Microsoft Authenticator app for passkeys and MFA, and scope conditional access policies to strengthen privileged accounts with phishing-resistant MFA.

Other tips included:

• Enabling Zero-hour auto purge (ZAP) in Office 365
• Configuring Microsoft Defender for Office 365 Safe Links
• Invest in ‘advanced’ anti-phishing solutions

"The companies that have a handle on AI governance (policies and procedures) well will be the ones that make safe AI use easy, risky AI use visible, and malicious activity hard to ignore. That means publishing a clear list of approved tools, blocking obvious lookalike domains and very recently registered domains can help stop this kind of threat," said Bruggeman.

"Monitoring suspicious downloads and sign-ins, and training employees on the AI-themed lures should also be done right now - don't think that generic phishing examples from five years ago are going to cut it today."

FOLLOW US ON SOCIAL MEDIA

In recent years, the traditional IT buying cycle has transformed. What was once a decision made largely at the discretion of CIOs, procurement, and operations is now significantly influenced by the chief information security officer (CISO).

The CISO has historically been a technical role, focusing on data security, cybersecurity, and overall risk management from an IT perspective. The modern CISO, however, has become central to business strategy and operational resilience, amidst growing cyber risks and increasing regulatory scrutiny. 

These shifting priorities and demands have led to CISOs becoming the linchpin between IT departments and the C-suite. However, this new pressure to boost business resilience, mitigate risks, and hold a central seat in business strategy decisions has led to widespread industry burnout.

For channel partners, this is a watershed moment as they now have the opportunity to engage with CISOs as strategic advisors, as well as product suppliers.

The CISO is the new buyer 

The expansion of the CISO’s role can be largely attributed to the sharp rise in cyber attacks over recent years, compounded by the rapid evolution of AI capabilities.

The UK government recently reported that almost half of businesses suffered a cybersecurity breach or attack within the last 12 months. And while AI readiness sits at the top of the C-suite agenda, uncertainty around infrastructure readiness is slowing down adoption, with many organizations revealing they struggle with data lineage, security, and governance. 

In light of these findings, it comes as no surprise that 69 percent of IT leaders are kept up at night by data security concerns, according to our research. Data is at the heart of all of these concerns: how it’s accessed, governed, and optimized. As a result, CISOs have become indispensable for ensuring continuous and proactive resilience and operational continuity, overseeing the intersection of security, innovation, and compliance. 

As data governance, lineage, and compliance become board-level priorities for enterprises, the CISO’s expertise is increasingly needed at the C-suite table, with many now reporting directly to the CEO.

For channel partners, this presents an opportunity to a trusted partner of CISOs who are navigating the new demands of their changing roles by offering solutions that signpost long-term compliance, operational continuity, and risk mitigation. Those who position themselves to address these new sets of priorities will ultimately build stronger long-term relationships with CISOs through offering personalized, practical guidance that will move the needle and help protect their organizations.

Regulation as a sales driver

While always slightly behind the newest threats, compliance standards around the world have been making strides to keep pace. This manifested in shifting away from static, periodical security scans and audits, to mandating secure-by-design solutions and continuous monitoring under acts like the EU AI Act, DORA, and NIS2. In the UK in particular, the upcoming Cyber Resilience Act will require organizations to ensure stricter visibility across their IT environments, which is a foundational requirement for enterprises gearing up for AI implementation. 

However, these new regulatory demands have created a widening confidence gap. Indeed, our research suggests that less than one-third of IT leaders have confidence that they will pass their next regulatory audit. As security becomes a consideration from the outset instead of an added feature at the end, CISOs command a larger budgetary influence in IT modernization initiatives, in order to avoid penalties for non-compliance. 

The global compliance market is estimated to reach $92.1 billion by 2033, presenting a clear market potential for channel partners. With CISOs now playing a key role in safeguarding reputation through risk mitigation, the channel has the opportunity to engage with this role as a strategic stakeholder, offering best-in-class solutions and tailored implementation guidance.

The channel can take this further still by positioning itself to offer region-specific expertise to CISOs, playing a key part in ensuring that their organizations are prepared to stay ahead of changing regulatory requirements.

Modernization through a security lens 

In fact, concerns over security block IT modernization strategy success the most, according to 41 percent of respondents to a Forrester survey we commissioned. This highlights a diminished risk appetite among enterprises. Instead, organizations are actively prioritising addressing vulnerabilities and improving visibility and governance across their entire IT landscapes.

This is especially important because any organization looking to implement and scale AI needs a high level of visibility and solid data governance to ensure compliant, high-quality, and secure data pipelines across the entire ecosystem.

Consequently, the way channel partners present value to CISOs has shifted, too. Customers now look for partners that can support their long-term strategies and improve their organizations’ cyber resilience and regulatory readiness, rather than just specific products or features. Partners should evaluate and present modernization solutions through a security lens in order to guide the CISO through the increased scope and responsibility that their role now entails. 

What’s more, the transformation of the CISO into a key figure guiding boardroom conversations is not complete. In fact, Gartner predicts (as cited in an IBM article) that regulatory pressure and attack surface expansion will lead to 45 percent of CISOs expanding their remit beyond cybersecurity by 2027.  

As enterprises are being challenged to modernize faster, strengthen cyber defenses, and govern data more effectively, they will continue to turn to the CISO for strategic guidance.

Channel partners that position solutions around the CISO’s core priorities will thus be better positioned overall to support the CISO in buying decisions, while building long-term trust by alleviating pressures facing the modern security leader.

Patch management is already challenging, but the issue is about to get worse as artificial intelligence (AI) tools such as Anthropic’s Claude Mythos Preview and OpenAI’s GPT-5.5 Cyber find and work out ways to exploit software flaws at scale.

So much so that the threat is being recognized at a national level. The UK National Cyber Security Center’s (NCSC)’s CTO Ollie Whitehouse has warned UK organizations to get ready for an expected surge in new software updates as the powerful new AI tools rapidly discover vulnerabilities.

Part of the attraction of frontier AI models such as Claude Mythos and GPT-5.5 is their ability to discover security issues more quickly. But their capabilities also put pressure on businesses to apply patches more rapidly than before.

It comes as vulnerability exploitation overtakes stolen credentials as the leading initial access vector in attacks, according to Verizon’s latest Data Breach Investigations report.

In this high-stakes environment, how can firms increase the speed at which they apply security fixes and mitigations?

The issue

The capabilities of frontier AI models are a stark challenge to traditional patching cycles. “The old assumption was that defenders had at least some time on their side to assess, prioritize , and patch before exploitation became realistic at any considerable scale,” says Rik Ferguson, vice president of threat intelligence at Forescout.

While that assumption has been weakening for several years, the threat has now become “supercharged”, he says.

The UK AI Security Institute has published benchmark data demonstrating how GPT-5.5 completed a 32-step simulated corporate attack chain end-to-end in two out of 10 runs. Meanwhile, Anthropic's Claude Mythos Preview performed the attack in three out of 10 cases. “Before Mythos, no AI model had completed that test at all,” Ferguson tells ITPro.

The benchmark data shows the need to address patching velocity is arriving faster than most organizations are ready for, according to Ferguson. “More fixes are coming. The question is whether your operating model is built to absorb them.”

Why AI tools add risk

Tools such as Mythos are designed for white-box, deep code analysis. “This is the kind that has the potential to surface undiscovered issues or vulnerabilities that signature-based scanning doesn’t reach,” according to Daniel Bechenea, security manager at Pentest-Tools.

When vendors run that continuously on their own products, the result is “more patches, shipped faster”, he says. “But that’s the upstream change. The downstream problem is, none of that vendor-side progress makes the organizations receiving those patches any faster at deploying them.”

Frontier AI tools will show up any gaps firms have in their patching processes. The immediate danger for most organizations is that existing weaknesses become “less forgiving, faster”, says Ferguson. “If you already struggle with asset visibility, patch prioritization, change windows, dependency mapping, or testing updates safely, AI-assisted vulnerability discovery makes those operational gaps more consequential.”

He says the pressure has shifted from "can we patch eventually?" to "can we absorb a much faster cadence of new fixes without breaking the business?"

It is as much a resilience problem as a security one, according to Ferguson. “And it can’t be solved by just buying a new tool. If the fundamental processes are not there, you have other work to do first."

The organizations most at risk are those that haven’t built the operational infrastructure – asset inventory, ownership mapping, evidence-based triage – to process findings at current volumes, Bechenea says.

And the problem will only get worse as more frontier AI models enter the cybersecurity market. Broadly, it will lead to faster vulnerability discovery and exploit development, as well as machine paradigm attack vectors, according to Ferguson. “These are attacks no human adversary would have designed, discovered by systems that don't think the way defenders were trained to anticipate. There will be far greater pressure on human-speed remediation processes,” he warns.

Ivan Milenkovic, VP risk technology EMEA at Qualys, concurs with this analysis. He thinks frontier AI will make individual timelines shorter, putting the standard “CVE release-patch before exploit” approach under “significant pressure”.

Prioritizing fixes

Rather than simply increasing the speed at which you patch, experts agree that prioritizing fixes is important. In 2025, there were more than 48,000 vulnerabilities discovered, and it will become an even greater issue with more powerful AI models in the mix, according to Milenkovic.

“The response to this can only be hyper-prioritization of issues, and fixing them – or at least the bulk – at machine speed,” he says.

The foundational security basics will also go a long way. The NCSC recommends measures beyond patching alone. Whitehouse advises focusing on “cyber security fundamentals” to raise resilience and to reduce the impact of breaches – including Cyber Essentials, or the Cyber Assessment Framework for organizations operating essential services.

AI tools can assist with specific parts of the workflow: For example, matching patches to affected assets, surfacing the highest-priority items, and flagging vulnerable components in your environment, says Bechenea.

Ferguson agrees that AI can “absolutely help”, inside governed boundaries. “Used properly, AI can improve vulnerability triage, code review, dependency analysis, test generation, and patch validation. It can help organizations decide what matters first and accelerate the surrounding work that usually creates delay.”

But to boost patching velocity, firms must ensure accurate asset inventory, dependency visibility, clearer prioritization based on exposure and business impact, tighter testing discipline, and pre-agreed decision paths for urgent updates, says Ferguson. “The organizations that patch faster successfully are the ones with fewer unknowns, fewer approval bottlenecks and better segmentation, so each patch carries manageable rather than existential operational risk.”

The goal is to reduce the time between knowing and doing, while “proactively managing the potential blast radius when you cannot patch immediately”, he advises.

The best responses will be based on understanding the business impact of any risk and how likely that is to be exploited, says Milenkovic. “That risk rating will change over time, based on what threat actors are doing and what other factors are in play.”

Security firms are warning that self-replicating malware known as Miasma has spread to 73 Microsoft GitHub repos across environments, including Microsoft Azure and Durable Task.

Miasma is a new and improved variant of Mini Shai-Hulud from the threat group TeamPCP, and, according to Cloudsmith, initially struck the @redhat-cloud-services npm namespace by compromising a Red Hat employee’s GitHub account.

"By pushing unreviewed orphan commits to internal repos, the threat actors injected a minimal workflow that requested GitHub’s OIDC tokens. This registry poisoning workflow in early June executed an obfuscated payload that published 32 malicious package versions to the npm registry," said the firm.

"Crucially, because it used legitimate OIDC tokens, the malicious releases carried valid SLSA provenance attestations. To standard registry scanners, the malicious updates were entirely indistinguishable from legitimate, routine code updates."

It's not known how many times the affected tools have been downloaded, but Microsoft said it's notified a 'small number' of customers who may have done so.

Under the hood of Miasma malware

What's special about the Miasma worm, said Cloudsmith, is that it doesn't exploit any software vulnerability in GitHub or npm, but instead exploits the underlying trust model of the modern engineering ecosystem.

Compromised developer credentials led to a legitimate GitHub OIDC token being requested, followed by a malicious build being published with valid SLSA provenance.

This ultimately led to conventional scanners seeing it as a routine trusted update.

On top of this, because Miasma generates a uniquely encrypted payload for each individual infection, traditional hash-based IOCs are functionally useless for broad detection, as the file signature changes with every single package version.

"While previous iterations of the Mini Shai-Hulud malware have focused purely on local secret scraping, the Miasma worm appears to have advanced data collectors specifically engineered for cloud identities in GCP and Azure," the researchers said.

"It attempts to harvest every cloud identity the infected developer machine and CI/CD runners have access to, proving a clear intent from the threat actors to leverage access away from the codebase and directly into live cloud environments."

How to protect your organization

If your company operates within the Azure or Red Hat ecosystems, Cloudsmith said to assume secrets exposure and rotate.

Miasma specifically hunts for developer credentials, meaning that everything on a compromised machine or CI/CD pipeline may have been been leaked.

"Developers are high-value targets because they sit at the intersection of source code, cloud infrastructure, AI platforms and production systems. Compromising a trusted package or development workflow can give attackers access that is far harder to obtain through traditional intrusion methods," commented Ilkka Turunen, field CTO at Sonatype.

With this incident having reached users of platforms such as Claude and Gemini, Turunen noted it shows how "interconnected modern software ecosystems have become" and should serve as a warning.

"An attack that begins with a seemingly insignificant open source package can quickly cascade across organizations, platforms and users," Turunen commented. "Organizations need to treat the software supply chain as part of their security perimeter. The attackers already do.”

FOLLOW US ON SOCIAL MEDIA

The National Cyber Security Centre (NCSC) has urged organizations to review their dependencies in light of an increasing number of supply chain attacks.

Recent attacks, the agency noted, have included maintainer account compromise, where attackers steal credentials or tokens that allow a malicious actor to update a trusted package.

Attackers are also taking over ownership of expired domains connected to package maintainers, or otherwise transferring ownership of a previously legitimate package.

Meanwhile, typosquatting is on the rise, with packages published using similar names to the genuine article, or with the misspelling of popular legitimate packages in the hope they are installed by mistake.

Threat actors are also using credentials or tokens stolen from a previous attack to access or modify additional packages.

These risks arise because one single application may rely on a large number of third-party packages – including libraries, frameworks, snippets, software development kits - some of which may not be entirely trustworthy.

Node.js, Rust and Python, for example, are unusually exposed as they have minimal standard libraries, boosting the use of third-party dependencies and delegation of basic functionalities, and leading to a heavy reliance on external registries.

Many of these components are retrieved automatically through continuous integration and continuous delivery (CI/CD) pipelines, often without human intervention.

"It is this combination of automation, trust and scale which means that malicious code introduced into a single package can spread rapidly across many organisations and services before detection," the NCSC warned.

The NCSC warned that threat groups are actively targeting developer environments, which are usually less tightly controlled than managed corporate devices, making it easier to compromise and steal the credentials of shared code or package registries.

A single malicious package can spread quickly across downstream software products and services. Indeed, the impact of compromising a lesser known, but critical, software component can have a significant and far-reaching impact for many organizations and systems.

The NCSC highlighted Node.js in particular, as its highly modular packages depend on many smaller components.

CI/CD threats are rising

Meanwhile, recent attacks have exploited the implicit trust in CI/CD and automation pipelines, where the automation of updates, installation, and execution of scripts and packages allows attackers to execute malicious code.

"For example, Node.js and Python support scripts that execute on installation, and allow a malicious package to be run immediately. Without human intervention or approval, the code can simply propagate," the NCSC warned.

Open publishing models increase exposure, with security controls for maintainer registry accounts not currently enforced by all registry providers.

Check your dependencies

The NCSC outlined a series of actions organizations are advised to take, including:

• Pause automatic dependency updates where compromise may be present
• Review and approve new updates, dependencies, or versions manually
• Rotate exposed or potentially exposed credentials
• Enforce MFA for developer and package registry accounts
• Use private or trusted registries where appropriate

"These attacks highlight the need to revisit how dependencies are introduced and managed, as part of a secure development lifecycle (SDLC)," the NCSC said.

"Whilst Node.js, Python and Rust are considered higher risk for these attacks, it’s important to be aware that other languages, tools, and package repositories are also at risk."

Developers should also make use of the Software Security Code of Practice, reviewing how dependencies are introduced and updated, avoiding automatically adopting new dependency versions without review, and striking a balance between deploying patches quickly and updating dependencies slowly.

This will help minimize the potential impact of compromise, according to the NCSC.

Elsewhere, they should also ensure deployments occur through controlled CI/CD pipelines rather than developer devices and store sensitive credentials securely, avoiding exposure on developer workstations.

"Modern software development has transformed how software is created, shared and reused – but recent attacks on these tools highlight the rapidly growing risks of using modern software ecosystems," the NCSC said.

"Whilst Node.js, Python and Rust are considered higher risk for these attacks, it’s important to be aware that other languages, tools, and package repositories are also at risk."

FOLLOW US ON SOCIAL MEDIA

Password management firm Dashlane said it has completed its investigation into an attack that allowed hackers to steal around 20 encrypted vaults.

The incident kicked off on Sunday, May 31, when a hacker launched an attack against a number of Dashlane user accounts by brute-forcing two-factor authentication (2FA) protections, allowing them to register new devices on existing accounts.

Because of the high volume of attempts on user accounts, Dashlane revealed its security controls automatically locked the accounts that were targeted by the attack.

However, the attackers were able to download a copy of the encrypted vaults of around 20 personal plan users, all of whom have now been notified, with some customers being prevented from adding new devices or logging in to their account with 2FA.

"Dashlane vault data cannot be accessed without the Master Password, and our vault encryption ensures that any attempts to gain access to the vault are statistically unlikely to succeed, even over a long period of time," said the firm.

"There is no evidence that Dashlane’s internal system has been impacted."

How the Dashlane attack unfolded

When a user enables an additional device, Dashlane verifies the identity of the account holder in a process that ends up sending a one-time six-digit token to the user’s registered email address.

For users who have enabled 2FA, a six-digit code generated by their authentication app is sent.

Once the user enters this code into the Dashlane application, Dashlane registers the device and downloads a copy of the encrypted vault to the device. The user can access this by entering the Master Password, which serves as the decryption key to the user vault.

"Without the Master Password, a user cannot access the items inside the vault. The vault encryption (Argon2 + AES-256-CBC + HMAC-SHA256) used by Dashlane ensures that any attempts to gain access to the vault are statistically unlikely to succeed, even over a long period of time," the company explained.

"Dashlane never stores Master Passwords or their derivatives on our servers in line with our zero-knowledge architecture."

New safeguards introduced

Dashlane said it has now deployed additional protections at the network level and within the product to increase the likelihood of detecting and filtering out malicious traffic.

Similarly, the firm will introduce additional layers of verification to the new device registration flow.

It also advises users to review the devices registered to their account and remove any that they don't recognize, and to enable 2FA on their account if they haven't already.

There's no need to change credentials or update the Master Password, said the firm, unless it's weak or easily guessed.

FOLLOW US ON SOCIAL MEDIA

Cyber criminals are using AI to increase the effectiveness of their attacks, according to new research from Anthropic, particularly in the later, more complex stages of their cyber operations.

In a study 832 accounts that were banned by Anthropic for malicious cyber activity between March 2025 and March 2026, the company found that 560 (67%) were using AI to write malware, with 7% using it to assist with lateral movement.

According to the AI firm, across that period attackers’ use of AI shifted from techniques to gain initial access towards actions carried out once they were inside.

For example, the use of AI for account discovery – identifying valid accounts inside a compromised environment – rose by 9%, while AI-assisted phishing fell by about the same amount.

"This suggests that attackers are increasingly applying AI deeper in the attack life cycle," the researchers said.

"These sorts of post-compromise techniques used to be restricted to actors with the technical knowledge to carry them out. Our investigation shows that AI can now be made to perform these activities on behalf of less sophisticated actors."

Lowering the bar for would-be hackers

Worryingly, Anthropic said that as AI is used to chain together many parts of the attack, it's getting harder to differentiate between high-risk and low-risk actors.

While the least-skilled actors in the dataset used about 16 distinct techniques on average, the most skilled used about 20 – not an enormous difference.

Similarly, there was no correlation between an actor's risk level and the specific platform used, whether Claude Code, an API, or a chat interface.

What does help distinguish higher-risk actors is where in the attack lifecycle they apply AI.

For example, they concentrate their use of AI on those techniques that require significant time, oversight, or real-time decision-making, such as account discovery, lateral movement, and privilege escalation, rather than just on tasks that allow them to gain initial access to the system.

They also design architectures that allow models to chain together discrete stages of a cyber attack and carry them out with minimal human input.

The researchers warn that the MITRE ATT&CK framework doesn't yet fully capture the tools and activities that make AI-enabled attackers so dangerous – such as using AI to orchestrate steps in the attack chain sequentially, make real-time decisions about what to do next, and execute without human intervention.

"Consider the state-sponsored cyber espionage operation we disrupted in November 2025. In that case, a malicious actor manipulated Claude Code into attempting to infiltrate targets around the world, with little human intervention. Mapping it against the MITRE ATT&CK framework shows that the actor used 30 techniques across 13 tactics, which was comparable to many medium-risk actors in our dataset," the researchers said.

"Clearly, focusing on the number of techniques this actor used underplays how dangerous they really were (by contrast, applying our risk-scoring methodology to this attack earns it the maximum risk score of 100)."

The company's now in talks with MITRE about how the ATT&CK framework might evolve to include the AI-enabled behaviors it's spotted.

FOLLOW US ON SOCIAL MEDIA

The global cyber threat landscape is undergoing a radical transformation, moving away from monolithic ransomware cartels toward highly volatile, fragmented splinter groups, a top UK police official has warned.

Speaking at Infosecurity Europe 2026, William Lyne, Head of Economic and Cybercrime at the Metropolitan Police Service, told IT and security leaders that the modern cyber crime ecosystem has evolved into a highly accessible space.

Lyne compared the underground landscape to a bar where threat actors can "get everything but a good drink."

"It felt like cyber threats were all quite stovepiped. You had hacktivists, you had hostile state actors," Lyne explained, reflecting on his early career. Today, however, those lines have blurred. "Those kind of stovepipes... no longer really exist."

Instead, Lyne described a blended ecosystem of products, goods, and services that has dramatically lowered the barrier to entry for prospective criminals.

This shift has been heavily accelerated by cryptocurrencies, which solved the traditional criminal bottleneck of "cashing out."

Previously, threat actors lost up to 75% of their profits navigating complex, expensive money-mule networks. Today, cryptocurrency allows them to realize illicit gains almost instantly and with very little risk.

Fragmentation and the 'post-trust' era

While massive international law enforcement operations have successfully dismantled groups like LockBit and disrupted phishing as a service (PhaaS) platforms, Lyne cautioned that the criminal underground is rapidly adapting.

"It's getting more diverse... [and] also much more fragmented," Lyne said. Following high-profile law enforcement crackdowns, cybercriminals have realized that operating as a massive, centralized brand or ransomware as a service scheme is "actually quite bad for business."

As a result, major ransomware operators are breaking off into smaller, independent factions.

This fragmentation is leading to a dangerous "post-trust" trend within the criminal ecosystem. Without the strict moderation and internal rules previously enforced by large cartel administrators, smaller threat actors are exhibiting more extreme, aggressive, and unpredictable behaviors.

The demographics of these attackers are also shifting. Lyne noted that the threat landscape is moving beyond traditional Russian-speaking hubs to include actors from Brazil, Türkiye, and English-speaking groups like the notorious Scattered Spider collective.

AI weaponizing hoarded data

Addressing the inevitable topic of AI, Lyne dispelled fears of autonomous systems launching end-to-end cyber attacks, but highlighted a pressing new risk for enterprise data privacy.

"These guys are generally not innovative," Lyne noted, explaining they only change their methods if they are “systematically earning less money... or they spy an opportunity to make more money."

Having stolen and hoarded petabytes of corporate data over the last decade, data that was rarely deleted even when victims paid the ransom, cyber criminals are now using AI tools to operationalize these massive "treasure troves" and mining historic datasets for new extortion and revenue streams.

Rewriting the law enforcement playbook

Faced with this agile, commoditized threat, the Met Police and its international partners are adopting aggressive new disruptive strategies.

"We can't arrest our way out of this problem," Lyne admitted, citing the jurisdictional complexities of cross-border cybercrime.

Instead, policing has shifted toward systemic disruption, psychological operations designed to undermine criminal trust, and targeting the foundational infrastructure of the cybercrime supply chain.

Crucially, this requires unprecedented collaboration with the private sector. Lyne emphasized that the Met Police is increasingly sharing intelligence with enterprise IT security teams and even naming industry partners who assist in operations on their site takedown pages.

"Ultimately, like, lots of these things just come down to trust," Lyne concluded, addressing the security professionals in the room.

"We want to have meaningful, both strategic and tactical collaboration with industry partners that we know hold some of the keys to... the challenges that we have in this space. The cultural change that we have undertaken, I think will continue so that we collaborate better moving forward."

FOLLOW US ON SOCIAL MEDIA

IT leaders and network defenders must stop treating cyber attacks as theoretical risks and start viewing them as acts of war, Ukraine’s former Minister of Foreign Affairs warned attendees at Infosecurity Europe 2026.

Speaking to a packed auditorium at ExCeL, Dr Dmytro Kuleba, whose own journey to the conference was delayed after his vehicle struck Russian missile debris en route to Warsaw, delivered a sobering keynote on the intersection of kinetic warfare, cybersecurity, and business continuity.

Drawing heavily on Ukraine’s defense in the wake of the Russian invasion, Kuleba detailed how the modern battlefield has seamlessly merged with enterprise IT environments, transforming everyday business software into deadly espionage tools.

The weaponization of the software supply chain

In a stark warning to technology buyers about supply chain security, Kuleba revealed how routine customer relationship management (CRM) systems used by small businesses were exploited by Russian intelligence to track and target Ukrainian citizens.

“What Russian security services are doing is they break into CRM systems of fitness clubs, salons, and the loyalty programs of supermarkets to track your movements,” Kuleba explained.

He noted that he was personally ordered by Ukraine’s security services to change his gym and barber to avoid being tracked.

According to Kuleba, this intelligence gathering was facilitated by the decades-long proliferation of Russian-made software in the Ukrainian market. The data collected from these seemingly innocuous business systems was subsequently used for blackmail and kidnapping.

“If even CRMs can be weaponized, basically any type of business, even the smallest one, will have to invest more in cyber security to protect itself from such a breach,” he warned, pointing to an accelerating shift toward sovereign and secure tech stacks. “Do not trust products made by your potential enemy.”

Business continuity and the “Mike Tyson” approach

Kuleba also urged CIOs to rethink their approach to disaster recovery and business continuity planning (BCP).

He referenced the devastating December 2023 cyber attack on Kyivstar, Ukraine’s largest mobile operator, which was brought offline after hackers compromised a single employee's account.

Despite the catastrophic breach, the telecom giant managed to completely restore and fence its systems within days.

For Kuleba, this level of recovery relies on deep environmental knowledge rather than rigid playbooks. Quoting boxer Mike Tyson’s famous adage that "everyone has a plan until they get punched in the face," Kuleba detailed how he prepared his own ministry for war in late 2021.

“We started planning... What if we wake up one day and don’t have access to our emails? What if we lose connection to our embassies abroad?” he said. The ultimate solution wasn't a complex procedural manual, but a deep audit of their architecture and the physical evacuation of their core servers to a safe location abroad.

"When it happens, you plan not to follow the plan, but to know your environment perfectly and develop instincts of survival in this environment," Kuleba advised.

"If you care for your company, you have to prepare for the worst. If it happens, you will instinctively be capable of winning, even though your initial plan will be ruined."

Redefining enterprise resilience

Ultimately, Kuleba challenged the security industry's conventional definition of resilience, arguing that it is no longer about bouncing back from a singular, isolated incident.

Whether dealing with a massive influx of disinformation, signal-jamming on the frontlines, or persistent network intrusions, defenders must accept a harsh new reality.

“Resilience is not about being prepared to repair destruction,” Kuleba concluded. “Resilience is your ability to keep repairing the networks as destruction becomes the new normal.”

FOLLOW US ON SOCIAL MEDIA

The rapid rise of flaw-spotting AI means companies need to bolster resilience plans to avoid becoming victims.

That's according to Commvault, which pointed to two key changes in security. Notably, advanced models are spotting a huge number of vulnerabilities — notably with the rise of frontier models like Anthropic Mythos and OpenAI's GPT-5.5 Cyber.

This increased level of automation is enabling threat actors to take advantage of exploits near-instantly, researchers warned. That collapse in the remediation window means resilience is no longer part of recovery, but an "operating requirement".

“AI models will continue to evolve that accelerate remediation timelines and require a new approach to readiness,” said Bill O’Connell, chief security officer (CSO) at Commvault.

O’Connell noted that resilience operations (ResOps) are now vital and an area that cannot be overlooked by IT leaders.

"ResOps gives organizations a way to continuously validate readiness, advance clean recoveries, restore systems with confidence, and build resilience into the way they operate."

CrowdStrike said earlier this year that AI is speeding up the pace of attacks, while Forescout said enterprises should be ready for an explosion in vulnerabilities. All of that means companies need to do more than simply patch in order to stay secure.

"Frontier models change the economics of vulnerability discovery. AI models will reveal exploitable vulnerabilities at such a fast pace, remediation programs must evolve,” said Nick Patience, VP and AI Practice Lead, Futurum Group.

"While a rigorous patching strategy remains critical, the key now is also making sure readiness, resilience, and clean recoveries are top priorities."

Cyber resilience in the AI era

To help enterprises stay ahead amid these challenges, Commvault recommended four key steps to set up a resilience operations framework, ensuring they can maintain business continuity through an attack, outage or AI driven disruption.

Risk evaluation

The first step is to evaluate the recovery risks, with IT and security assessing how well their current plans will hold up against faster flaw spotting and exploitation cycles caused by AI.

Commvault advised looking beyond backups and asking "harder questions", such as whether critical systems can be restored cleanly and if recovery environments are isolated from compromised production systems.

Similarly, IT and security teams are advised to ensure recovery plans have been mapped to key dependencies.

Isolation is key

After that audit, Commvault said the second step was to isolate recovery to ensure critical data remains secure and backed up to support remediation efforts.

"Maintain immutable, isolated copies of critical data and workloads, separated from production identity, network, and management planes," the company advised.

"These copies help provide a clean fallback when patching or when remediation cannot keep pace."

Beyond that, enterprises should assume that recovery time objectives set before the advent of AI will no longer hold true, and reconsider them against new attack scenarios.

Identify priorities

The third step is to prioritize any systems that are business critical, identifying those that are required for the business to function, be it identity platforms, billing systems, or cloud services.

Then, set out which order they should be recovered. Don't forget to include new dependencies such as data pipelines, model repositories, and agentic workflows.

Automation can bridge gaps

Lastly, organizations should automate where they can, according to Commvault. This could include automated threat scanning or recovery orchestration and restoration.

Regular testing of recovery plans is also critical, the company noted, which can be supported through automation. This is a vital area, researchers warned, largely due to the pace of change brought about by AI.

"Organizations that embrace this four-step process will be better suited to take advantage of rapidly evolving AI models while also mitigating the risks,” Patience added.

FOLLOW US ON SOCIAL MEDIA

Ransomware is still a booming business, according to new research from Rapid7 Labs. So much so that cyber criminal gangs are outperforming major companies.

Analysis from the cybersecurity firm found ransomware groups made an estimated $529.2 million in the first quarter of this year, with total revenues up by 39% year-on-year.

That's a better performance than FTSE 350 companies have managed in the same period, not one of which showed year-on-year revenue growth of over 30% during the quarter.

A number of major cyber crime outfits are profiting from the boom, and it’s been a particularly good year for the Qilin ransomware group. Rapid7 researchers noted the group made an estimated $193 million between July 2025 and March 2026.

The Gentleman group, meanwhile, made an estimated $52 million over the same period.

“Ransomware groups are not the isolated, hooded hacking crews in dark rooms," said Thom Langford, CTO EMEA at Rapid7. "Instead, many resemble highly efficient businesses generating revenue growth that would make legitimate organizations envious.”

Booming ransomware revenues

One reason for the booming revenues is the rise of initial access brokers, which has lowered the barriers to entry by shifting cyber crime from technically specialized malware development to a mature underground marketplace.

Access, tooling, and full attack services are now commercially available to almost anyone. Modern cyber crime operations involve distributed networks of specialists handling initial access, malware and ransom negotiations, and working like legitimate supply chains.

Servers, leak sites, and negotiation portals can be quickly rebuilt after disruption, while law enforcement takedowns take longer to coordinate and execute.

“The problem is they are demonstrating, very publicly, that ransomware can be a successful criminal enterprise, and ironically, in some ways, they’re more resilient than businesses themselves,” said Langford.

“Removing one group, one server, or one piece of infrastructure rarely collapses the wider operation because the ecosystem is designed to keep functioning around the damage."

Battling continued ransomware threats

Rapid7 said organizations should prioritize identifying and reducing exposed attack surfaces on a continuous basis, focusing on misconfigurations, isolated assets, and internet-facing vulnerabilities.

These are all commonly exploited in initial access brokerage markets, the study noted.

Elsewhere, security teams should leverage threat intelligence more proactively to map adversary behavior patterns, infrastructure, tooling, and access pathways.

Notably, researchers said defenses should shift toward preventing credential and access compromise at source. This includes implementation of stronger identity controls, enforcement of least privilege rules, and monitoring for early indicators of credential resale or misuse in underground ecosystems.

“To give ransomware groups the economic crash they deserve, we need to shift to earlier visibility and earlier action," said Langford.

"That means businesses understanding exposure, reducing attack surface, tightening identity controls, and using threat intelligence to intervene earlier in the chain before ransomware becomes an outcome rather than a possibility.”

FOLLOW US ON SOCIAL MEDIA

Identity Access Management (IAM) is a key building block to successful risk management at a time when ID theft is a key route for threat actors into company networks.

Since 2020, successful cyber breaches leveraging identity theft have become widespread, with recent examples such as the Visa, Marks & Spencer, Jaguar, and Harrods cases illustrating how even well-resourced companies are not immune to these types of attacks.

These breaches were linked to the Scattered Spider group, which has undergone a merger with another prolific cybercriminal group known as ShinyHunters. In just four months, the new group has successfully targeted and infiltrated multiple targets across the US and Europe, including the March 2026 breach of the European Commission, resulting in a 350GB data leak.

The anatomy of an identity attack

In these Identity attacks, Social Engineering is the primary initial vector, with cyber criminals leveraging Vishing and Phishing attack vectors to bypass SSO and multi-factor authentication (MFA) identity access controls. Attackers will often masquerade as internal IT, calling users on their work or personal phones to re-enroll or reset their IAM credentials, then send a modified account reset link to bypass non-phish-resistant MFA.

With credentials successfully hijacked, the attacker can then replay the MFA token to access SaaS resources and exfiltrate corporate data for the extortion phase.

The Social Engineering vector is popular as it offers an easy way to understand a company’s internal structure by leveraging social media. Employee names are usually linked to job titles, which reveal potential access privileges. Personal posts, interests, and commentary give insight into effective tactics for acquiring personal data through phishing, and, as corporate email addresses follow well-known conventions, they are easy to determine.

With identity being a pillar of cybersecurity but also a key attack vector, there are some key capabilities that IAM should deliver.

Rethinking IAM for a more complex threat landscape

Identity access management is not a one-size-fits-all solution. Customer environments and business objectives determine which identity controls will be the most effective.

Nevertheless, there are key capabilities an IAM solution should provide:

1. Coverage based on thorough integration within the corporate environment ensures a ‘single source of truth’, allowing visibility over the whole network – including legacy systems.
2. Correlation of login data used to identify potential anomalies. The more complex the environment, the more important this capability. Automated analysis can flag potential issues to be investigated manually as a second step, to uncover more details.
3. Reporting that enables pertinent and concise alerts to be raised by the IAM solution and follow a clear escalation path to ensure key stakeholders have actionable intelligence for decision-making.

Out of these capabilities, it’s the correlation element that is most important for the early detection of potential breaches in IAM integrity.

To improve the chance of early detection, it’s more effective to focus on looking for anomalies within the environment. These could be related to the user identity behavior, such as “Impossible Travel”, the user identity “location” represented by changes in IP address, or the service identities in the environment spiking in activity during off-hours.

There are several strategies that organizations can adopt to identify anomalous sign-ins without disrupting user experience; these fall under the concept of Risk-Based Authentication (RBA). Organizations can implement User and Entity Behavior Analytics (UEBA), which creates a profile of user behavior and can trigger a biometric or MFA check if a user activity falls outside of the baseline of the usual profile.

Conditional Access is another option, triggering authentication when a user activity exceeds a defined risk score threshold. Integrating FIDO2 passkeys, either software-based or hardware tokens, with one of the above RBA methods will greatly improve the efficacy of RBA by eliminating 90% of the common “anomalous sign-in” flags generated by password guessing or phishing.

IAM as a managed service opportunity

With threats on the rise and limited in-house cybersecurity expertise, companies of all sizes increasingly rely on managed cybersecurity services to strengthen and maintain their security posture. The IT channel is in a privileged position to deliver tailored and effective solutions incorporating IAM as an essential element of corporate cyber-resilience. But what should a robust IAM managed service include?

A true managed identity service should include an MFA or Passkey (FIDO2) capability, allow for customized policy management, and be able to deliver identity services to both users and non-human systems. It should also be capable of risk analysis powered by machine learning and AI, and deliver workflow orchestration. The service should be continuously reviewed and updated to keep up with the fast-evolving threat landscape.

For partners building IAM-managed services, it is recommended that they first conduct housekeeping in their own environment. Supply chain compromise is one of the top concerns in 2026, and partners must be able to show that their own environments are secure.

Secondly, if you have access to multiple vendors, you should standardize your solution stack. Ideally, you would have two core identity platforms, with one likely to be Microsoft Entra ID.

Thirdly, you should develop a comprehensive onboarding blueprint. The success of the service will depend on a positive customer onboarding experience, minimizing any outages in the process to ensure business continuity.

Ultimately, identity is no longer just an administrative layer. It is central to how organizations defend their environments.

As attackers increasingly target credentials, access pathways, and identity stores, businesses need IAM strategies that combine visibility, detection, and strong authentication. For partners, the opportunity lies not simply in selling another security tool, but in helping customers build a more resilient and adaptive approach to identity-led risk.

IBM and its subsidiary Red Hat are pumping $5 billion into improving the security of open-source projects.

Project Lightwell is designed as an enterprise clearinghouse for open source software, with a new AI-driven model for securing the software supply chain.

The idea is to use advanced AI capabilities, offered through commercial subscriptions, to validate and test fixes across a huge volume of open source code. Enterprises will be able to integrate secure patches directly into their existing software supply chains, say the firms, with enterprise-grade validation and lifecycle management.

They can report and resolve vulnerabilities, receive patches optimized for production environments, spanning both Red Hat offerings and independent community code, and share fixes upstream so that open source communities can include them in long-term maintenance.

And all this will be backed by a team of more than 20,000 engineers working across upstream and enterprise environments. The focus will be on upstream maintenance alongside open source community leaders; high-volume, AI-assisted vulnerability review, triage, and prioritization; and secure patch development, dependency hardening, and release engineering.

"Open source is the backbone of today's digital economy and the foundation of modern AI, and we are at an inflection point in how it is built, secured, and scaled," said Arvind Krishna, chairman and CEO of IBM.

"With Project Lightwell, IBM and Red Hat are helping define a new industry model, one that brings together AI, engineering expertise, and trusted collaboration, to secure open source software at its source and across the entire supply chain. This is about strengthening trust in the systems that power business, government, and society."

IBM and Red Hat are already working with a group of early adopters on Project Lightwell, including Bank of America, BNY, Citi, Goldman Sachs, JPMorganChase, Mastercard, Morgan Stanley, Royal Bank of Canada, State Street, Visa, and Wells Fargo.

"The real-world insights from these initial deployments will actively shape how vulnerabilities are identified, validated, and remediated at scale across complex software supply chains," the firms said.

More than nine-in-ten Fortune 500 companies rely on open source software - but security is an ever-present problem. Sonatype identified 454,648 malicious open source packages in 2025, up 67% on the previous year, with one state-linked group alone tied to more than 800 malicious packages.

Meanwhile, according to Black Duck, 86% of codebases contain open source vulnerabilities, with 81% of those classified as high or critical risk, up from 74% in the previous year.

"Most enterprises cannot keep up with the volume, complexity, and speed of risk. AI-driven vulnerability discovery is accelerating both the volume and speed of CVE creation, compounding an already unsustainable remediation gap," said IBM.

"Project Lightwell delivers validated fixes to the specific open source versions organizations already run. By combining large-scale engineering, AI, and a coordinated clearinghouse model, it enables organizations to move from detection to remediation without disrupting stability, certification, or compliance requirements."

Cybersecurity professionals are less likely to trust a boss who's never been through the mill of managing a major security incident.

Data from antivirus vendor Sophos suggests that CISOs have a one-in-four chance of losing their jobs after an attack. But new research from ISC2 shows that three-quarters of security professionals reckon leaders are more credible if they've already led their organization through a major cyber incident – regardless of how things turned out. Just 9% disagreed.

Overall, the survey revealed that the most trusted security leaders are those who create confidence through transparency, consistency, and an ability to align security priorities with business outcomes. Those who can keep calm and carry on, demonstrating decisive leadership under pressure, are far more likely to earn lasting credibility with their teams and across the enterprise.

Unfortunately, though, cybersecurity bosses don't generally seem to be managing this.

Only 34% of cybersecurity professionals said they were very confident in their current cybersecurity upper leadership, with 15% extremely confident. Three-in-ten said they had moderate confidence, 15% were only slightly confident, and 6% said they had no confidence in their cybersecurity leaders at all.

Security staff are particularly keen on leaders who can communicate risk to senior leadership and boards, with 95% of respondents reckoning this as very important.

Other big pluses included a strategic and long-term cybersecurity vision, along with the ability to effectively work with senior leadership and boards to secure budget, and being transparent about decisions and actions.

Decision-making under pressure, building and leading high-performing teams, and technical cybersecurity expertise were all very important to more than eight-in-ten – more so than actual technical cybersecurity expertise, at 75%.

"The most important trait in a cybersecurity leader is the ability to align security strategy with business goals while earning trust through clear judgment, communication, and accountability," noted one respondent.

Bosses wanting to earn their staff's respect, said ISC2, need to be transparent about risks, priorities, and challenges. "Teams and executives are more likely to trust leaders who provide realistic assessments rather than overly optimistic narratives," the researchers said.

Keeping calm and carrying on in high-pressure incidents or periods of change also boosts a security leader's reputation, while there's much greater trust when leaders manage to create an environment where teams feel supported, heard, and accountable.

Strong cybersecurity leaders invest time in understanding business objectives and collaborating across departments, helping position security as an enabler rather than a blocker.

"For leaders who now find themselves in an environment where cybersecurity risk impacts every part of the organization, it is the ones who communicate clearly, empower their teams and demonstrate calm, decisive leadership under pressure that are far more likely to earn lasting credibility with their teams and across the enterprise," the researchers said.

"Ultimately, the most successful cybersecurity leaders are not simply those who protect systems and data, but those who create trust in their leadership when it matters most."

Hackers posing as IT experts are showing up in person at law firms, the FBI has warned.

In the past, the Silent Ransom Group (SRG), also known as Luna Moth, Chatty Spider, and UNC3753, sent phishing emails purportedly charging small 'subscription fees'. To cancel the fake subscription, the victim was instructed to call the threat actor, who then emailed the victim a link to download remote access software.

Now, though, the group is using phone calls and phishing emails to pose as IT support, gaining access to the victims' computers and exfiltrating data.

And while this is often done through legitimate remote access tools, the group has also been sending individuals in person to the victim company's location to gain physical access to machines.

"This is a pretty natural evolution of extortion operations. We spent years building detections around malware and exploits, and now attackers are shifting toward social engineering, trusted tooling, and physical access," commented Gabrielle Hempel, security operations strategist at Exabeam.

"Physical security fell by the wayside when organizations began to move their data to the cloud, but if your security model assumes that the threat actor is always on the other side of the internet, you have a problem."

The group's first step is to either directly call or send phishing emails urging employees to call 'IT support'. While on the phone, the SRG actor directs the employee to grant access to a remote desktop session.

If that attempt fails, though, SRG sends a threat actor to the victim's location to gain access and insert a storage device into the victim's computer. The hacker tells the victim they need to image the device or create a backup file to address potential impacts from the phishing email.

Once they've got access to the victim's device, they minimally escalate privileges and quickly pivot to data exfiltration without encryption, using Windows Secure Copy ( WinSCP) or a hidden or renamed version of 'Rclone'.

"SRG actors use the exfiltrated victim data to extort the victim by sending a ransom email threatening to sell or post the data online," the FBI said. "SRG actors also call employees or clients of a victim company to pressure the victim to begin ransom negotiations."

While SRG has hit companies in a number of sectors, including the insurance, finance, and healthcare industries, it's consistently been targeting US-based law firms since spring 2023.

"The group is leaning into trust by posing as IT support, walking employees through remote access, then moving quickly to steal data before anyone realizes something is wrong," warned Nick Tausek, lead security automation architect at Swimlane.

"That makes this especially dangerous for law firms. These environments hold sensitive client records, privileged communications, financial details, and case information. If that data is stolen, the damage does not stop at the victim organization. Clients can be pressured, legal strategies can be exposed, and employees can become targets for follow-up scams."

The UK is building a national cyber defence shield powered by AI, the first to be publicly announced in the world.

That's according to a speech from GCHQ head Anne Keast-Butler, who said AI was an "unstoppable force with great opportunity" that comes with risks.

The system she described remained a plan that the security agency hoped to have up and running within the next five years, according to reports, and would use AI agents to spot threats against critical national infrastructure, including airlines and telecoms.

"In the past few months, GCHQ has developed the blueprint for a new national cyber defence capability will hardwire cutting-edge agentic AI into machine-speed cyber defence," Keast-Butler said.

"And as we draw on decades of expertise in machine learning to reimagine cyber security, we're also embedding frontier AI deeper into our operations – responsibly and ethically – to enhance algorithms, translate foreign languages, and find needles in haystacks faster than ever before," she added.

Few other details about the plans were revealed.

Last year, major British companies suffered serious outages after a series of cyber incidents, including one against Jaguar Land Rover that was deemed the "single most financially damaging cyber event to ever hit the UK" and another that cost retailer M&S nearly £400m.

Hackers have increasingly turned to AI for their attacks, with AI now deemed a "standard part of the attacker toolkit", according to one industry expert, while CrowdStrike said the technology was accelerating the expansion of enterprise attack surfaces at an alarming speed.

Future tech threats

She said AI efforts were necessary for the UK and its allies to stay ahead, and there was a narrowing window to keep the lead, in particular when it comes to technology and war.

"Tech companies are releasing AI-driven innovations at a remarkable pace, with untold consequences, as algorithms are weaponised often just below the threshold of traditional warfare," said Keast-Butler. "And China is now a tech superpower with sophisticated cyber, intelligence, and military capabilities."

She added: "One thing is clear: technology and data are no longer just tools; they are transformational forces."

Alongside AI, she pointed to quantum computing, calling for companies to be ready for the security implications, notably the ability to crack existing encryption techniques.

"So we must protect our most critical systems from future quantum attacks," Keast-Butler said. "This will take time – and that's why we're calling on businesses to act now to meet the timelines set out by NCSC."

The National Cyber Security Centre last year laid out a ten-year timeline for the transition to quantum-resistant encryption methods, known as post-quantum cryptography – but 90% of companies aren't prepared, according to one survey.

And she noted the rise of space technologies, noting that ten thousand new objects had been launched into space in the three years since she took over as GCHQ director.

Teamwork to counter threats

The UK and its allies, be it Europe, Five Eyes or NATO, need to work together to counter the danger of China or Russia taking the lead in any of these technologies – and both are investing heavily.

She admitted that raised issues for digital sovereignty, an increasingly hot topic amid geopolitical concerns presented by the dominance of Big Tech firms from the US and Chinese success in AI and hardware.

"Some strive to stay safe by keeping data within their borders, shutting out foreign IT," she said. "But that doesn't work."

The EU has started to tender for sovereign cloud services for EU institutions, and France is making its own software to replace US-made productivity solutions, for example. But Keast-Butler argued for a different route, saying sovereignty need not mean "made in the UK" but control over supply chains, data, and dependencies.

"It's about backing great British science and innovation, established tech companies, and academic excellence – whilst not limiting our ability to harness the best of the world's technology," she said.

A new threat group, The Gentlemen, has become one of the most active ransomware operators, accounting for 10% of all attacks and second only to the notorious Qilin.

Despite only having emerged in July last year, The Gentlemen has quickly evolved into a highly operational RaaS group, according to the NTT, using advanced tooling and proxy infrastructure to accelerate attacks and improve stealth.

With a level of technical maturity that would normally be associated with more established ransomware groups, the researchers believe that the group consists of experienced actors with potential ties to other ransomware ecosystems.

The group's targeting remains focused on industrial organizations, the information technology sector, and some consumer spaces, with notable victims including Synergy France, UK Electronics, and Equity Life.

In terms of target geography, meanwhile, The Gentlemen largely extorts organizations in Europe, with the UK and Germany among the most heavily targeted countries.

Its affiliates are increasingly leveraging SystemBC malware, a proxy and backdoor tool often used in human-operated ransomware attacks, to establish covert tunnelling, evade detection, and support rapid lateral movement across enterprise environments.

The group's rapid growth so far this year, combined with its sophisticated proxy infrastructure and obfuscation techniques, means organizations should expect faster intrusion cycles and reduced dwell times before encryption deployment, NTT said.

"The rise of groups like The Gentlemen demonstrates how affiliates are now combining shared tooling, stealth infrastructure, and repeatable intrusion methods to accelerate attacks at scale," said Matt Hull, VP of cyber intelligence and response at NCC Group.

"Techniques such as covert tunnelling and rapid domain-wide deployment are shrinking the window that defenders have to detect and respond before encryption occurs."

According to NTT, there were 748 ransomware listings worldwide during April, representing a 7% fall from the figure for March. However, ransomware activity in 2026 has been operating at a higher baseline than much of 2025, as the ransomware-as-a-service (RaaS) ecosystem expands and matures.

Claude Mythos – the large language model reportedly capable of autonomously identifying vulnerabilities and developing exploit chains – has yet to make its mark, thanks to restricted access, controlled testing environments, and questions around operational effectiveness at scale.

"Developments around AI models such as Claude Mythos suggest AI-assisted vulnerability discovery and exploitation could further compress attacker timelines in the future," said Hull. "However, the industry should remain cautious about overstating current capabilities, particularly where testing has been limited to controlled environments."

The report also highlighted several geopolitical developments likely to influence cyber activity in the coming months, including China's expanded supply chain security regulations, which consolidate and extend existing controls on import and export activities.

Meanwhile, the strategic significance of NASA's Artemis program is motivating China and other nations to carry out espionage, IP theft activities, and potentially even destructive attacks.

"Numerous other well-resourced countries (and private companies) are pursuing high-stakes interests dependent on the domain of space; including but not exclusive to India, Japan, Israel, South Korea, UAE, Russia, Iran, and North Korea," the researchers warned. "Defenders should avoid being too narrow in their assessments of potential threats."

The UK and Australia have announced plans to extend their cooperation on AI security and safety.

The two plan to share insights on AI capabilities, carry out research into emerging risks, and work together to develop international best practice for testing and evaluating AI systems. They plan to carry out joint research, including, they said, new approaches to the measurement, testing and management of risks.

"Australia and the UK have always worked closely to keep our people safe – and that partnership matters more than ever in the age of AI. This technology is moving fast, and so are the risks that come with it – particularly in areas like cybersecurity. No country can tackle that alone," said UK AI minister Kanishka Narayan.

"By working hand in hand with trusted partners such as Australia, we can stay ahead of the risks, strengthen our defences and make sure AI is used to improve lives for Brits and Aussies alike."

The new Memorandum of Understanding (MoU) between the two countries is based on a new partnership between the UK AI Security Institute and a newly-created Australian AI Safety Institute. This, said the government, will monitor, test, and share information on emerging AI capabilities, risks, and harms, making sure that Australia has the capability to keep up with developments in AI technologies. And the new deal will see the two institutes working closely, providing for staff exchanges to strengthen day-to-day collaboration.

"This MoU brings Australia's AI Safety Institute and the UK AI Security Institute together to share expertise, identify risks early and support safeguards that help prevent harm," said Australian minister for industry and innovation and minister for science Tim Ayres.

"Collaboration with our closest partners is how we make sure AI is safe and secure while we continue to explore the opportunities it presents."

AISI already has similar agreements with AI research bodies across the world's major economies, developed through the International Network for Advanced AI Measurement, Evaluation, and Science, and through its bilateral partnerships.

And this deal builds on a 2024 MoU between the UK and Australia on online safety and security covering illegal content, child safety, age assurance, technology-facilitated gender-based violence, and the harms of generative AI.

It committed the two countries to in-person dialogue, coordinated bilateral and multilateral engagement, regulatory engagement, shared research projects, and working with industry to address safety challenges.

"This agreement with the UK reflects our shared commitment to ensuring artificial intelligence is developed and deployed safely, securely and responsibly," said Australia's assistant minister for science, technology and the digital economy, Dr Andrew Charlton.

"By strengthening cooperation with democracies, we are combining world-class expertise to understand emerging risks and keep pace with rapid technological change."

At the end of 2021, a now infamous vulnerability was found in the Apache Log4j open-source logging library. It led to a race against time, with teams struggling to identify which apps and services were affected and apply the patch before attackers could exploit the bug.

As the risk of similar software supply chain threats multiplies, a software bill of materials (SBOM) can help. Essentially an inventory of open source and third party components, an SBOM is now part of regulations including the US Executive Order on Cybersecurity, the EU Cyber Resilience Act (CRA) and the EU Network and Information Systems 2 Directive (NIS2).

How do SBOMs work, and does your business need one?

Why firms need an SBOM

Most modern applications include open source and third party components. As a result, a large share of software risk comes from dependencies that teams did not write themselves.

“When a serious issue like Log4j appears, the first question is simple: Where are we exposed?” says Ilkka Turunen, field CTO at Sonatype. “If you cannot answer this question quickly, you have a real problem.”

Log4j is “the prime example of why SBOMs can be indispensable”, says Dana Simberkoff, chief risk, privacy and information security officer at AvePoint. “Because the library is very broadly used in consumer and business-facing software applications, the attack on Log4j had catastrophic consequences for a wide range of applications across regions and sectors.”

An SBOM might have prevented the worst effects of the attack by allowing IT professionals to more easily expose and patch the vulnerability that triggered it, according to Simberkoff.

Visibility into software supply chains

As software supply chain threats become more complex, organizations need complete visibility into the components that make up their applications. This visibility is “critical for security”, says Crystal Morin, senior cybersecurity strategist at Sysdig.

She points out that modern vulnerabilities are being weaponized much more quickly. “So when a vulnerability is discovered in a software component, SBOMs allow organizations to quickly determine whether they are affected and see where that component is in use.”

It comes at a time when the urgency to fix bugs is increasing. In April, Mozilla said it used Anthropic’s Claude Mythos Preview to find and fix 271 bugs in Firefox. Anthropic has argued that teams could find many hundreds of vulnerabilities by applying current frontier models.

As AI super-charges vulnerability discovery, it creates a need for much better visibility into software, says Turunen.

An SBOM performs this task by offering security, engineering and legal teams a shared understanding of “which components are in use, where they came from, and what needs attention when something goes wrong”, says Turunen. “In practice, this means less guesswork, faster coordination, and a better chance of keeping up as both software complexity and vulnerability discovery speed increase.”

Who needs an SBOM

Regulators are now mandating an SBOM across critical sectors. In the US, Executive Order 14028 requires stronger software supply chain security for federal systems, including the use of SBOMs. Meanwhile, agencies such as the National Institute of Standards and Technology (NIST) and the National Telecommunications and Information Administration (NTIA) have published guidance for federal entities and their software suppliers.

The EU CRA, set for full enforcement in December 2027, requires SBOMs for products with digital components sold in the EU. In the UK, the government has introduced SBOM guidance through its Software Security Code of Practice.

The UK’s National Cyber Security Centre (NCSC) encourages organizations to adopt SBOMs to boost software supply chain security. This is despite the fact they are not yet mandated in domestic regulation, says Ritchie Perry, electronics engineer at ByteSnap Design. “The direction of travel is the same: more transparency over what’s in the software you ship, and better evidence that you’re tracking known vulnerabilities.”

Even if you are not in a regulated market yet, there is a good chance your larger customers are – and they will start flowing SBOM requirements down the supply chain, Perry adds.

Industries where SBOMs are most critical are “where the stakes are highest”, says Simberkoff. “For example, software products that interact with critical infrastructure, government agencies and services, and emergency services.”

How to create an SBOM

A typical SBOM consists of supplier name, component name, version, dependency relationship, the author of the SBOM, the time the data was added to the SBOM, and any known risk related to security, legal, and quality. “Together, this information helps secure the intricate and vast systems that constitute software supply chains,” according to Turunen.

Creating an SBOM should be an “automated, integrated part of your software development lifecycle”, says Simberkoff. “It should not be treated as a one-time task because software changes. Every time software is patched or changed, your SBOM should automatically update.”

Most teams will lean on tooling, says Perry. He points out that products such as Black Duck, Snyk, Mend and FOSSA can scan codebases and generate SBOMs.

Meanwhile, platforms such as GitHub and GitLab include SBOM support from their dependency graphs. “These tools are helpful, but not perfect, so it is worth spot checking the output, particularly around high-risk components,” adds Perry.

SBOMs also must be analyzed for risk using threat intelligence and policy rules, monitoring each vulnerability proactively in case it needs to be replaced with a more secure alternative.

To integrate compliance and security throughout business processes, SBOMs must be exported and shared in standard formats for use across teams and external stakeholders, says Turunen.

Turunen also has a warning: “Even if an organization knows what its developers are using, that doesn’t mean their knowledge extends to vendors. This means not just using SBOMs, but requiring them from vendors as well.’’

It’s also important to realize that simply generating an SBOM is not enough. Organizations must take proactive steps to ensure SBOMs are “living assets used to drive better security decisions”, says Turunen.

Keeping SBOMs current is a top priority, concurs Perry. “Components change, dependencies update and new CVEs appear constantly. An SBOM that reflects last year’s build is far less useful than one generated from yesterday’s.”

The FBI has warned about an attack against Microsoft systems using yet another Phishing as a Service platform called Kali365 that can gain access without intercepting user credentials.

The Kali365 platform was first spotted last month, and the US agency said it wanted to alert users of the threat caused by the Phishing as a Service (PhaaS) platform – a technique that's on the rise, with PhaaS kits increasing in sophistication and becoming more popular among rookie hackers.

"Kali365 has primarily been distributed via Telegram, enabling cyber threat actors to obtain Microsoft 365 access tokens and bypass multi-factor authentication (MFA) protocols without intercepting the user's credentials," the US Federal Bureau of Investigation (FBI) said via a public service announcement.

Hackers with a subscription can use Kali365 to snag "OAuth" tokens that allow persistent access to Microsoft 365 environments.

While that's concerning on its own, the FBI noted that the Kali365 platform makes it easier for hackers to have success targeting such systems. "Kali365 lowers the barrier of entry, providing less-technical attackers access to AI-generated phishing lures, automated campaign templates, real-time targeted individual/entity tracking dashboards, and OAuth token capture capabilities," the statement said.

Microsoft has yet to respond to ITPro's request for comment on the attacks. This isn't the first time a PhaaS platform has targeted the tech giant. Earlier this year, Microsoft teamed up with security agencies to take down the Tycoon 2FA PhaaS platform that was also targeting Microsoft 365 logins, while Microsoft worked with Cloudflare to take down a similar PhaaS system also targeting Microsoft credentials last year.

How the Kali365 attack works

The attack will target victims via an email that's pretending to be a cloud productivity or document sharing service. "This phishing email contains a device code with instructions to visit a legitimate Microsoft verification page and enter the code," the FBI explained.

Such emails will be specifically targeted to individuals or organisations. If fooled, the user will be directed to a real Microsoft page. When the authentication code is pasted into the page, the victim will be authorizing access – not to their own devices, but those managed by the attacker.

That's possible because the attacker captures OAuth access and refresh tokens, letting them take over the targeted person's Microsoft 365 account.

"The attacker can now access Microsoft 365 services such as Outlook, Teams, and OneDrive without needing a password or completing any additional MFA challenges," the FBI added.

What to do

As ever, users can avoid becoming a victim by not clicking links in unexpected emails – this may be an advanced Phishing as a Service platform, but the attack vector remains a standard phishing email with a dodgy link.

At a corporate level, to help protect against this style of attack, the FBI advised restricting the amount of codes that are used for authentication, such as creating a conditional access policy to block device code flow for everyone, only allowing limited exceptions for necessary processes.

That said, the FBI noted it may make sense to audit how such codes are used now to ensure no legitimate use cases are disrupted.

Beyond that, companies can implement policies that block the transfer of authentication from computers to mobile devices, and exclude emergency access accounts.

The FBI added that anyone impacted by Kali365 – be it phishing emails, suspicious logins, or spotting unauthorized devices – should file a report with the Internet Crime Complaint Centre (IC3).

GitHub has confirmed that around 3,800 internal repositories have been breached, after a developer unwittingly installed a malicious VS Code extension.

The Microsoft-owned code repository and DevOps platform said the breach was detected on Monday, but that the activity involved exfiltration of GitHub-internal repositories only.

"We have no evidence of impact to customer information stored outside of GitHub's internal repositories, such as our customers' own enterprises, organizations, and repositories," said the firm's chief information security officer, Alexis Wales.

"Some of GitHub's internal repositories contain information from customers, for example, excerpts of support interactions. If any impact is discovered, we will notify customers via established incident response and notification channels."

GitHub said it started rotating critical secrets as soon as it discovered the breach, with the highest-impact credentials prioritized first. It is now analyzing logs, validating secret rotation, and monitoring its infrastructure for any follow-on activity, it said, promising a fuller report once it's finished its investigation.

GitHub hasn't explicitly named the attacker, but made reference to a claim by the TeamPCP hacker group that it had accessed around 3,800 repositories, saying that the number was consistent with its investigation so far.

TeamPCP, which first appeared late last year, is the group linked to the Mini Shai-Hulud worm, and carries out supply chain attacks by stealing CI/CD credentials and using them to publish infected versions of further packages.

The group has reportedly not asked for a ransom for the GitHub data, but is offering the stolen data for sale for $50,000, saying that if it doesn't receive an offer, it will leak it for free.

"This is another reminder that developers are now permanent targets in software supply chain attacks. TeamPCP has shown how a motivated attacker can move through the tools developers trust every day – open source packages, extensions, accounts, and credentials – rather than trying to break in through the front door," said Ilkka Turunen, Field CTO at Sonatype.

"Combined with the acceleration we're already seeing from AI-assisted vulnerability discovery, the window between compromise and exploitation is collapsing. The old assumption was that defenders would have time to identify, prioritize, and respond. That margin is disappearing."

The news came just a day after the Nx Console VS Code extension, which has 2.2 million installs, was briefly backdoored, with the malicious version collecting credentials silently when a developer opened a workspace. The issue was handled swiftly, with the extension pulled within 18 minutes on the VS Code Marketplace and 36 minutes on Open VSX.

"The community's ability to catch and remove malicious packages is real. For extensions with millions of installs, it's also insufficient," commented Shaun Brown technical product marketer at Aikido Security.

"Caught in 18 minutes and prevented exposure are not the same thing. Minimum package and extension ages are the best way to protect your devices from similar attacks today."

Wasabi Technologies has announced a series of updates aimed at accelerating growth across its EMEA partner ecosystem, as demand grows for cyber resilience and AI-ready storage services.

The cloud storage provider wants to strengthen its channel proposition through new ransomware recovery capabilities, deeper technology integrations, and simplified management tools designed to help partners deliver secure and scalable storage services.

The additions come amid increasing pressure on organizations to secure the rising volumes of data tied to AI deployments and cloud applications. Wasabi's 2026 Global Cloud Storage Index revealed that 42% of European organizations experienced a cyber attack that resulted in a loss of access to their cloud data, while 92% are now backing up production AI data and applications.

According to the company, the increasing number of corporate AI deployments has heightened the importance of data security as part of wider cyber resilience strategies.

"Cyber resilience has moved from an IT priority to a business imperative, and our partners are leading this change of mindset across EMEA," commented Kevin Dunn, Wasabi's vice president and general manager for EMEA.

Strengthening enterprise-grade security

To address these growing concerns, Wasabi has expanded its cyber resilience capabilities with Covert Copy, a patent-pending ransomware recovery tool that creates an invisible and immutable copy of critical data designed to remain inaccessible during an attack.

The vendor has also expanded its partnership with Commvault to deliver a joint data protection and cloud storage solution, while also enhancing its integration with Synology to support hybrid backup and storage deployments that combine on-premises infrastructure and scalable cloud storage.

According to Wasabi, these integrations will simplify the deployment process for MSPs and resellers, as well as bolster cyber resilience strategies through risk mitigation, immutable back-up capabilities, and cost-efficient data management.

"With innovation like Covert Copy, deeper ecosystem integrations with long-term partners like Commvault and Synology, and flexible consumption models, we're enabling partners to deliver secure, high-performance storage solutions that protect critical data while giving customers the cost control and simplicity they need to scale," Dunn added.

In a bid to further drive partner growth, Wasabi has also introduced Wasabi Account Control Manager Self-Activation, a new capability that equips MSPs with free, multi-tenant management with one-click onboarding to help reduce time-to-market and drive operational efficiency improvements.

Accelerating partner growth

According to Wasabi, its partner-focused strategy continues to drive momentum across the region, with partners leveraging its predictable pricing model, zero egress fees, and high-performance architecture to win competitive deals in fast-growing areas such as AI and data protection.

Neil Brosnan, vendor alliance director at Exclusive Networks, said the company's partnership with Wasabi is helping channel partners capitalize on growing demand for secure and scalable data management solutions as AI adoption continues at pace.

"Our partnership with Wasabi is built on a shared belief that cyber resilience starts with data – how it's stored, protected, and governed," he explained. "As organisations adopt AI and modern applications, the volume and value of data is exploding, bringing new regulatory pressures and a broader attack surface."

The cost of unplanned downtime for large firms has shot up by 50% in just two years, with Global 2000 companies losing around $600 billion per year.

In its new report, The Hidden Costs of Downtime, Cisco unit Splunk has found that downtime now costs organizations an average of $15,000 per minute and nearly $95 million in lost revenue annually – twice as much as in 2024.

The most severe cost is publicly disclosing a data breach, with just over seven-in-ten technology executives rating it as very or prohibitively disruptive – more than three times as many as in 2024.

Alongside the obvious costs, the survey found, organizations see an average 3.4% drop in stock price following a downtime event. And there's also an effect on customers, with 47% admitting customers are often the first to detect service degradation or outages, and, as a result, 81% said this had ultimately cost them customers. Worse, nearly 20% of marketing professionals report that it takes an entire quarter to recover brand health following remediation.

"It's the direct impact on your customers and employees that can turn a technical problem into a total business crisis," said Kamal Hathi, SVP and GM at Splunk.

Meanwhile, with threat actors now calculating ransoms based on a victim's specific downtime costs, payouts have nearly tripled since 2024, reaching $40 million on average.

"Watching companies suffer nine-figure losses makes the abstract risk of a cyberattack feel real and immediate," commented Peter Sprenger, Splunk field CTO.

Thanks to regulations like the EU's GDPR and Digital Operational Resilience Act (DORA) and new standards for financial institutions, fines have escalated significantly over the last three years. They're now costing an average of $51 million per organization, with 57% of technology executives now viewing these penalties as very or prohibitively disruptive.

And there's also the issue of operational drag, with 89% of tech leaders citing the need for large numbers of personnel to fix issues. Nine-in-ten tech leaders said they've seen demand for customer support rise, with 76% of finance and 74% of marketing executives also feeling the pressure.

One thing that doesn't help is that, according to 36% of security leaders, downtime is often or very often misclassified as an IT issue, and only 38% saying they can consistently identify the root cause of a downtime incident.

And SaaS and other third-party application issues are on the rise, nearly tripling since 2024, with 56% of security leaders now experiencing these issues often or very often.

The best tactic, said Splunk, is to introduce strong AI-driven incident response and observability capabilities. Three-quarters of organisations identified as 'AI Workflow and Triage Experts' were able to avoid the need to publicly disclose a data breach last year, compared with just 54% of non-experts, and were also nearly three times more likely to report that they have never lost customers due to downtime.

"Downtime is inevitable; prolonged disruption is not," said Hathi. "The most resilient organisations are not the ones with the most tools or the biggest vision for AI. They are the ones that align technology with business outcomes, empower people with context, and design systems that bend, but do not break, under pressure."

Cyber attacks are typically associated with data theft and extortion, but another threat can cause just as much damage. As geopolitical tensions rise across the globe, state-sponsored adversaries are preferring to hide in systems, going unnoticed for months or years after the initial compromise.

To perform these so-called living off the land (LotL) attacks, attackers are weaponizing legitimate software and infrastructure to lie in wait. This trend is seeing tactics shift away from data breaches to more sophisticated espionage and disruptive operations, according to a new report from Cloudflare.

Why are LotL attacks growing right now, and how should firms respond to this threat?

Living off the land attacks: long-term campaigns

Cyber attacks usually take advantage of security weaknesses. However, living off the land attacks are different: They are growing in response to organizations strengthening their overall cybersecurity posture, Tony Fergusson, CISO in residence at Zscaler tells ITPro. “Organizations have made significant progress in their ability to detect threats and patch systems more effectively. Consequently, adversaries are being forced to be more stealthy to exploit data, and they’re doing this by leveraging legitimate tools and processes.”

With living off the land attacks, attackers deliberately avoid drawing attention to themselves by using existing and trusted tools and websites, rather than exploiting a zero-day flaw or introducing malware, says Fergusson. “They stay under the radar, blending in seamlessly with legitimate user activity, and mimic everyday operations so their presence goes unnoticed.”

Cloudflare’s 2026 threat report describes a shift away from brute force entry towards high-trust exploitation, with adversaries actively targeting legitimate SaaS, IaaS, and PaaS tools such as Google Calendar, Dropbox and GitHub to camouflage malicious actions within normal enterprise activity.

This isn’t surprising, says Razvan Ionescu, head of offensive security services at Pentest-Tools.com. He describes how his team “consistently finds that organizations have invested heavily in signature-based detection and perimeter controls”. Yet the monitoring of legitimate administrative tooling, endpoint management platforms, cloud management consoles and scripting environments “remains thin”.

State-sponsored and highly-targeted

Living off the land attacks suit a certain type of adversary. The technique is especially attractive to “state-sponsored and highly-targeted threat actors”, according to Dana Simberkoff, chief risk privacy and information security officer at AvePoint.

Rather than seeking immediate financial gain, attackers are aiming for espionage, strategic positioning and in some cases, preparation for future disruption. “Living off the land tactics allow these adversaries to maintain access over long periods without drawing attention,” Simberkoff explains.

Living off the land attacks allow nation states to collect strategic intelligence across diplomatic, military, economic, or technological targets, says Tracey Hannan-Jones, consulting director for information security at UBDS Digital. “By using pre-positioning, attackers gain access to critical systems, so disruption can be triggered during geopolitical tensions.”

Supply chain attacks, seeing adversaries compromising vendors to reach downstream targets, are “easy leverage”, warns Hannan-Jones.

Cloudflare’s report tracked four primary nation state adversaries over the past year: Russia, China, North Korea, and Iran. Each group approaches living off the land attacks differently based on its operational goals, Ionescu tells ITPro.

For example, China appears to have shifted from bulk data theft towards targeting legitimate cloud infrastructure for longer-term pre-positioning, with groups such as FrumpyToad using Google Calendar for command-and-control communication.

“The goal is to create a resilient architecture that remains nearly invisible to standard perimeter defences,” says Ionescu. “Rather than trying to exfiltrate data today, these attackers are establishing persistent footholds now to use during a future geopolitical event.”

Living off the land attacks: businesses most at risk

Certain businesses are more at risk from living off the land attacks than others – especially in critical sectors and those holding data valuable to nation state adversaries.

Organizations with complex digital environments are particularly exposed, says Simberkoff. “Cloud-first enterprises, regulated industries, critical infrastructure providers and companies embedded in large supply chains are at risk.”

The more identities, integrations and third party connections an organization has, the more opportunity attackers have to hide, warns Simberkoff. “Risk also increases for organizations that are strategically interesting to nation state actors, whether because of the data they hold or the role they play in a broader ecosystem.”

Government and defense are prime targets for living off the land attacks. “State actors look at pursuing intelligence and influence, accessing and stealing sensitive data, policy insight and information of geopolitical value, so they can use it against them,” says Hannan-Jones.

Stealthy with technology

Rapidly developing technology such as AI is adding to the risk, allowing attackers to perform increasingly stealthy attacks.

The current shift is subtle. AI is making attacks “more refined”, says Simberkoff. “Instead of fully autonomous attacks, we’re seeing AI used to support reconnaissance, targeting and decision making. This helps attackers understand environments faster and choose techniques that look the most legitimate.”

The result is activity that increasingly resembles normal administrative behavior, which makes detection much more difficult, she warns.

Attackers can use AI to rapidly analyze public information such as organization charts, job postings, technical blogs, vendor documentation and leaked credentials and infer likely tech stacks and access paths, according to Hannan-Jones. “This improves the precision of initial access attempts and reduces the need for noisy trial-and-error.”

How to protect your business from living off the land attacks

Living off the land attacks are a concern, but there are some steps firms can take to boost their security.

Rather than trying to prevent compromise entirely, Simberkoff recommends focusing on “detecting misuse and limiting impact”. She advocates strong identity governance, least privilege access and “detailed logging of administrative activity”.

Ionescu underscores the importance of “understanding your own blast radius”. “Before asking what you’d detect, ask what an attacker with compromised admin credentials to your endpoint management platform, your identity provider or your cloud management console could do,” Ionescu advises. Most organizations haven’t mapped that explicitly.”

The second priority is closing the gap between “what your monitoring covers” and “where attackers actually operate”, says Ionescu. “Effective reconnaissance from an attacker’s perspective focuses on maintaining OPSEC and blending into normal traffic patterns and avoiding detection at the earliest stages of the kill chain. Your detection logic needs to match that: Anomaly detection on administrative actions, not just signature matching on known bad payloads.”

Robust incident response is also key. Protecting your firm from living off the land attacks requires building operational playbooks for “quiet compromise”, says Hannan-Jones. “Many organizations will have playbooks for ransomware, but very few are prepared for stealthy pre-positioning. Define what ‘suspicious admin activity’ looks like in your environment and create response runbooks for identity compromise, token theft and privileged account misuse.”

Vulnerability management is widely regarded as the cornerstone of cybersecurity. It’s built on the premise that if organizations can identify and prioritize their vulnerabilities, then they are far better placed to manage risk effectively.

A practice that’s been around for years, it’s allowed security teams to create some semblance of order in what can, at times, feel like a world of chaos.

But if vulnerability management is about identifying, assessing, and fixing known weaknesses, what about the things that are beyond its scope? Between cloud sprawl, remote work, third-party tools, and evolving threats, today’s modern attack surface is bigger and more dynamic than ever.

Recent incidents, such as the exposure of Anthropic’s Claude Mythos preview through third-party access and predictable infrastructure patterns, show how risk increasingly sits outside traditional boundaries.

As a result, traditional vulnerability management – which was built for a slower, more predictable world – simply can’t keep up.

Faced with this growing realization, the answer for many businesses and organizations lies in exposure management.

The increasing relevance of exposure management

Instead of focusing only on known vulnerabilities, exposure management surveys the entire attack surface – from misconfigurations and external threats to identities and unknown assets – to pinpoint where risk is most acute.

Add in the fact that it also focuses on the vulnerabilities that are most likely to be exploited and cause real damage, and it’s easy to see why exposure management as an approach is gaining traction.

What’s more, not only does it enable IT teams to prioritize the most critical issues, it also provides them with the details and context they need to remediate vulnerabilities and weak configurations.

That includes finding, assessing, and prioritizing risks across everything connected to an organization's IT environment, including cloud systems, internal servers, third-party services, and even forgotten or unused web assets.

Exposure management is the next logical step beyond vulnerability management, giving organizations a single, continuous view of risk and the ability to turn insight into fast, targeted remediation.

Too much data, not enough action

Part of the reason for this shift is that security teams are now inundated with data. With tens of thousands of new vulnerabilities disclosed each year, the challenge is not just the scale of the problem, but the pace at which it continues to grow.

A study published by Forrester Consulting on behalf of Google Cloud in 2025 – Threat Intelligence Benchmark – found that 61% of those surveyed said they were overwhelmed by the sheer volume of information, with a similar number (59%) admitting that they struggled to work out which threats were real and relevant.

Studies such as this are useful because they help shine a light on a problem that many businesses and organizations face. But it doesn’t tell the whole story.

You can think of it as having a detailed map of every pothole on the road, but never actually getting around to fixing them. Or like knowing exactly where the weaknesses are in your defenses, but leaving them exposed.

Remediation-first exposure management

This is where a remediation-first approach to exposure management comes in. Instead of merely identifying and prioritizing issues, the emphasis here is on fixing problems.

In a sense, it turns a more traditional approach on its head. Instead of being something that is put on a list to be dealt with at “some point in the future”, the remediation is dealt with there and then.

This has implications for how security and IT functions operate. It requires closer coordination and a clearer understanding of which vulnerabilities are genuinely exploitable. And it also requires teams to act in a more coordinated fashion, especially across large, distributed environments.

AI and automation also start to play a more practical role. By helping IT teams to prioritize the most relevant issues and carry out remediation more quickly, it adds a whole new level of efficiency to the process, whether through patching, updates, or policy enforcement.

Crucially, this is also the direction the industry is heading next – towards autonomous IT – moving beyond visibility to solutions that actively reduce risk in practice. Accenture’s Technology Vision 2025, for example, highlights a shift to “AI… acting autonomously on behalf of people” across enterprise systems and operations.

How MSPs and resellers can deliver and demonstrate value

For MSPs and resellers, this creates a clear opportunity to move beyond simply providing visibility and, instead, focus on helping customers prioritize and fix exposure as part of their everyday operations.

In practical terms, this means taking a remediation-first approach to managing systems, endpoints, networks, and applications, while also chipping away at the backlog of unresolved issues.

This shift is not going to happen overnight. Many businesses and organizations are still heavily invested in traditional approaches and need to be convinced of the benefits, both in terms of reducing risk and improving operational efficiency.

That is where MSPs and resellers have a critical role to play, not only as technology providers, but as strategic advisors who help customers understand what must change, why it matters, and how to act.

Improving cyber risk management and meeting compliance goals requires more than visibility; it demands decisive action. Anthropic’s Claude Mythos preview has now accelerated the speed of relevance and execution, making the decisions organisations face far more urgent and reinforcing the need to move toward Autonomous IT as a matter of priority.

The main risk from hospital cyber incidents is no longer data breaches or IT disruption – it's direct threats to care delivery.

According to a Black Book Research survey of 284 European hospital cybersecurity buyers, 82% rate their 2026 cyber attack concern as very high or extreme, while 74% believe their organization is likely or highly likely to face a major cyber event this year.

And, the researchers found, attacks are no longer viewed primarily as privacy events, compliance events, or IT disruptions – but as threats to the delivery of care.

"Europe's hospitals are operating in one of the most complex cyber-risk environments in the world: nationally connected health systems, public-sector capacity pressure, cross-border supplier ecosystems, aging infrastructure, accelerated cloud migration, strict regulatory accountability, and clinical operations that cannot go offline," said Doug Brown, founder of Black Book Research.

"Attackers know the pressure points. They are not only targeting data; they are targeting authentication, availability, recovery windows, third-party dependencies, and the fragile digital workflows that move patients through emergency departments, labs, imaging, pharmacy, theatres, ICUs, and discharge."

As a result, European hospital cybersecurity buying has shifted sharply from breach prevention toward clinical continuity. Two-thirds are investing in identity, IAM, PAM, SSO failover and break-glass access, and 57% in ransomware recovery, immutable backup, and read-only clinical access.

Just over half are looking to network segmentation, zero trust, and ZTNA, 46% to incident-response retainers and crisis-response services, and 45% to third-party supplier and vendor cyber-risk management. Meanwhile, 37% are investing in medical device/IoMT security, and 29% in cyber range, downtime simulation, and resilience exercise services.

However, while 78% of survey respondents said their board receives general cybersecurity risk updates, only 31% receive cyber-resilience metrics tied to clinical continuity.

Only a quarter reported a full clinical downtime simulation within the past 12 months, and 32% said their organization had never conducted a full clinical downtime simulation, had only completed tabletop activity, or did not know when the last exercise occurred.

Worryingly, while 59% of respondents said they were confident that their hospitals could operate safely for 24 hours without core Electronic Health Record (EHR) access, that figure fell to 32% at 48 hours and just 14% at 72 hours.

"The 72-hour number should disturb every hospital board and ministry-level health technology leader in Europe. A hospital that can improvise through the first day of downtime is not necessarily resilient," said Brown.

"By day two and day three, medication reconciliation, laboratory turnaround, radiology workflow, identity access, pharmacy verification, transfer coordination, discharge planning, and backlog reconciliation become patient-safety risks. Cyber resilience is now an operational medicine issue."

The health sector is an increasingly popular target for cyber criminals, thanks to its critical nature. And many attacks have led to problems delivering patient care, including a 2024 ransomware attack on NHS pathology provider Synnovis, and, more recently, an attack on medical technology firm Stryker being described by the firm as 'destructive, not ransomware'.

"In Europe, the cyber battleground has moved from the server room to the bedside," said Brown.

In the rulebook of cyber crime, a stolen credential is a gift that keeps on giving. Instead of breaking in, an attacker with access to stolen credentials can simply log in. No alarms get triggered as no intrusion is detected. Besides the immediate danger of identity theft, credential-based cyber attacks can expose organizations to distributed denial of service (DDoS). This type of attack causes systems to crash, bringing mission-critical business operations to a standstill.

Put simply, one stolen password equals one whole organization’s potential undoing. Even more alarming is the rampant rise of information-stealing malware, also called infostealers.

In 2025 alone, Flashpoint identified more than 11.1 million infostealer-infected devices. Subsequently, threat actors harvested a little over 3 billion credentials and cloud tokens, underscoring the increased exposure of authentication data.

A safeguard that’s always on, however, can act as a strong deterrent against credential-based attacks. Traditional authentication methods, including two-factor authentication, verify a user’s identity only at login – a weak link frequently exploited in brute-force attacks.

Continuous authentication helps address this security gap by prompting re-authentication based on device location, user activity, and security policies. No user, device, or application is considered inherently safe. Instead, trust is earned continuously.

How continuous authentication works

Once integrated into an application, a continuous authentication system constantly computes an authentication score that reflects an active user’s authenticity. If the score drops below a set threshold or confidence level, the system puts forth additional verification. This could be, among other things, a fingerprint or temporary access code. If the verification fails, access to the application or device is revoked. Multiple behavioral cues influence the authentication score, including blink rate and typing speed, making infiltration a less stealthy endeavor.

The mode and frequency of authentication may vary depending on the service provider. Certain tools may also allow security screening at specific instances of user activity.

While effective on its own, continuous authentication can be paired with other digital safeguards for a more layered approach to cybersecurity. Just-in-time (JIT) access is a case in point. Rather than providing continuous or standing access, the security mechanism grants privileges only when needed and for a limited duration. By strictly controlling access timing and scope, JIT access shrinks the window of opportunity for threat actors.

“Cyber criminals are most often breaking in without breaking anything – capitalizing on identity gaps overflowing from complex hybrid cloud environments that offer attackers multiple access points,” stated Mark Hughes, global managing partner of cybersecurity services at IBM, in a 2025 threat report.

“Businesses need to shift away from an ad-hoc prevention mindset and focus on proactive measures such as modernizing authentication management, plugging multi-factor authentication holes and conducting real-time threat hunting to uncover hidden threats before they expose sensitive data.”

Continuous authentication as part of identity and access management also helps meet safety standards set by the European Union’s General Data Protection Regulation (GDPR) and the US National Institute of Standards and Technology (NIST) Cybersecurity Framework (CSF).

In the event of an account hijack, continuous authentication can still protect sensitive data by continuously monitoring user behavior. Limiting what an attacker can access even after a break-in is a key benefit of continuous authentication. Suspicious activity can trigger immediate session termination.

Continuous authentication and zero trust

The use of continuous authentication remains limited across organizations, but this could soon change. That’s because continuous authentication is a key pillar of zero trust and zero trust network access (ZTNA), as the technical backbone for its ‘never trust, always verify’ approach. As organizations strengthen their zero trust model, they will need to implement continuous authentication throughout their enterprise environment.

In hybrid and remote work environments, the security and resilience gains of continuous authentication are also clear to see. With the right safeguards and protocols, continuous authentication can improve security outcomes without discounting regulatory requirements.

The use of continuous authentication finds further support in guidance from the National Institute of Standards and Technology (NIST), which encourages organizations to screen passwords against lists of known compromised and commonly used credentials.

When a match is detected, appropriate responses may include prompting an immediate password reset, alerting security teams, or triggering automated remediation. Minimizing reliance on periodic security reviews, this approach facilitates continuous credential integrity.

With a reported sale of 300,000 AI chatbot credentials on the dark web, it’s safe to say identity theft has become a low-effort crime. Drawing attention to the ease with which such attacks can now be carried out.

Incorporating threat intelligence into continuous authentication can result in a more cohesive and adaptive security posture by narrowing the interval between breach and discovery. While continuous authentication helps detect abnormal user behavior during active sessions, threat intelligence adds contextual awareness of emerging risks and known attack patterns. Together, they can enhance both response speed and detection accuracy.

Enterprises have experienced a sharp increase in the number of identity-related breaches over the last year, according to two new studies.

According to new research from Sophos, 71% of organizations suffered at least one identity-related breach across 2025, with organizations reporting an average of three separate incidents and 5% reporting six or more.

The main consequences of an identity-related breach are data theft (49%), ransomware (48%), and financial theft (47%), the study found. Indeed, two-thirds of ransomware attacks were carried out this way, with serious financial consequences.

Sophos noted that the mean recovery costs associated with ransomware attacks reached $1.64 million, with a median of $750,000. Nearly three-quarters (73%) of those affected faced costs of $250,000 or more.

“Identity has become the primary attack surface in modern cybersecurity, and this data shows most organizations are losing ground,” said Ross McKerchar, chief information security officer at Sophos.

“The non-human identity problem is particularly urgent. AI agents are being granted privileges faster than security teams can track them, and organizations that fail to get ahead of this will find it an increasingly costly gap to close.”

Enterprises have a visibility problem

Visibility is a critical weakness, according to Sophos, with only a quarter of organizations continually monitoring for unusual login attempts, and more than half checking every three months or less.

Detection, meanwhile, is equally poor. Around 14% of breached organizations were unable to detect and stop their most significant identity attack before damage was done.

A key factor for many identity breach victims lay in compliance, according to the Sophos study. Among those that found compliance requirements challenging, 82.4% had suffered a breach – a full 14 percentage points higher than those with less difficulty with compliance.

UK firms grappling with identity security

In the UK specifically, enterprises are contending with similar challenges. According to Palo Alto Networks' Identity Security Landscape Report 2026, machine identities now outnumber humans 100 to one, creating serious identity security risks.

82% of organizations expect to see the number of machine identities rise over the next 12 months, the study noted, and 90% expect to see a sharp increase in AI identities.

More than one-third (34%) of AI agents and 37% of machine identities have access to their organization’s data, which may include sensitive information such as financial records or high value systems.

Palo Alto Networks noted that only 51% of UK organizations are using behavioral monitoring for their autonomous AI agents.

Identity security has become a key focus – and pain point – for many enterprises since the advent of agentic AI. With agents given deep access to sensitive data sources, risks are amplified and the potential for data leakage is now a leading concern for IT and security leaders alike.

Fragmented tools create blind spots

Fragmented identity security systems and tools are also causing problems with regard to visibility, according to eight-in-ten UK firms. Respondents to Palo Alto Networks’ survey said disparate tools are impacting or delaying their ability to detect and respond to identity-related threats.

As a result, 83% of UK organizations have experienced an identity-related breach, while 74% have fallen victim to at least three in the last 12 months.

“The explosion of machine identities represents a fundamental shift in the enterprise attack surface. With AI-driven identities projected to continue accelerating in the next year, organizations are facing a reality where identity complexity is rapidly outpacing traditional security controls," said Rich Turner, Palo Alto Networks' senior vice president EMEA.

"The fact that 83% of organisations have suffered an identity-related breach in the UK - and 91% in EMEA more broadly - proves that as AI agents gain more access to sensitive data, security leaders must move beyond manual processes. To close the gap, organizations must embrace end-to-end automation and unified governance. Otherwise, the risks of expanding AI and machine identities will only continue to intensify.”

FOLLOW US ON SOCIAL MEDIA

AI models are getting better at handling more complex security tasks, doubling results in one benchmark in the last few months – and that's before the arrival of security-focused models, notably Anthropic's Claude Mythos and OpenAI's GPT-5.5.

That's according to the UK's AI Security Institute (AISI), which tracks the potential impact of AI on the security industry and efforts to protect organisations, and found newer models had doubled the length of cyber tasks they could complete in just 4.7 months – much faster than expected.

"In February 2026, we internally estimated that the length of cyber tasks AI models could complete had doubled every 4.7 months since late 2024 – already an acceleration from our November 2025 estimate of 8 months," the organisation said in a blog post. "Since then, AISI reported on two new models, Claude Mythos Preview and GPT-5.5, which substantially exceeded both doubling rate trends."

AISI added: "It is unclear whether this represents a new, faster trend."

That follows the release of Claude Mythos, which sparked concerns that companies wouldn't be able to keep up with AI security, as well as GPT-5.5 Cyber last week, with OpenAI releasing the security focused model in a limited preview with access only to security professionals, amid fears that generative AI was accelerating a security arms race. Indeed, Forescout VP of security intelligence Rik Ferguson last week said AI tools are now "a standard part of the attacker toolkit."

How the AISA tests

These results are based on a time-horizon benchmark, which tracks the success rate of AI models on tasks of different lengths based on how long a human expert would take on the same task. For example, one set of tests includes reverse engineering and web exploits in self-contained setups. AISA is looking for a model to succeed 80% of the time to be considered capable of doing a task of a certain length.

AISI admits the time horizon benchmark is imperfect. "They are inexact predictors of performance; AI struggles with some tasks humans do quickly, and easily completes others that humans find hard," the blog post noted. "However, we use this type of benchmark because it offers a measure of AI autonomy from which we can draw trends."

Plus, the tests only include some capabilities that would be necessary to run a real-world attack. Alongside that, AISI limits the models to 2.5m tokens to maintain comparability across results.

But AISI said that the 2.5m token cap limits the success of models, as without that cap, the "success rates are so high that time horizons become impossible to calculate." But the organisation also added that its own tests are now too short, meaning it's not possible to reveal at what point model reliability would start to fail on a longer task; the longest task is 12 hours.

"No single benchmark result should be read as a precise measure of AI capability," the post noted, adding: "Regardless, the direction of change and rapid growth have been consistent across the models, methodological choices, and independent data we examined."

New evaluation methods were in development, the AISI added.

What this means for security

The AISI said it was unclear how AI's pace of progress would continue, or how the technology's capabilities would work against real-world systems. But the agency said it was clear that AI was bringing opportunities and risks.

"The time to invest in strong security baselines is now," the AISI post warned. "Frontier AI can strengthen attackers as well as defenders, and there is a critical window to build resilience."

That was echoed by Palo Alto Networks this week, with CTO Lee Klarich warning that AI cyberattacks would become the new normal in the next few months. "This impending vulnerability deluge demands urgency," he wrote.

Attacks on operational technology (OT) are surging, according to new research, with industrial organizations the biggest target of ransomware in every single month over the last year.

According to a new analysis from cyber security firm NCC Group, in the 12 months to March 2026, industrial organizations experienced 2,073 ransomware attacks, accounting for 30% of all ransomware activity.

Manufacturers of capital goods such as machine equipment and infrastructure were particularly hard-hit, accounting for 1,192 attacks. Within this industry, machinery was the most-targeted sub-sector, with 442 attacks, followed by construction and engineering with 394.

“Our data shows that many organizations continue to prioritize IT security while underestimating the exposure of their operational environments," said Ray Robinson, OT director at NCC Group.

"When OT systems are disrupted, the impact goes far beyond data loss - production can halt, essential services can be disrupted, and in some cases, lives can be put at risk.”

Governments worldwide are growing increasingly concerned about the issue. In the UK, Network and Information Systems (NIS) Regulations require operators of essential services to put proportionate technical and organizational measures in place to manage cyber risk across both IT and OT environments.

Meanwhile, the Cybersecurity Act and sector-specific guidance cover OT governance, incident reporting, resilience, and supply-chain security.

“Regulators are increasingly clear that OT environments fall within scope of cyber resilience obligations, particularly where systems support essential services or public safety," said Katarina Sommer, global head of government affairs and analyst relations at NCC Group.

"Organizations that focus compliance efforts solely on IT risk are exposing themselves to operational, regulatory and safety consequences, so it’s key that organizations treat OT risks in the same way they approach IT security.”

Earlier this year, the National Cyber Security Centre (NCSC), along with US, Australian, Canadian, and European authorities, issued a new guide for OT owners and operators aimed at helping them integrate 12 security considerations into their procurement processes.

These include making sure that the product allows for security and safety logging, has strong authentication controls, protects data, is configured in a secure way by default, and is supported by established vulnerability management processes by the manufacturer.

“As cyber attackers increasingly target operational technology around the world, it has never been more vital for critical infrastructure operators to ensure security is baked into the systems they use," said Jonathon Ellison, NCSC director of national resilience and future technology.

State-backed hackers targeting operational technology

Many attacks on OT systems come from nation state-affiliated actors, with the US Office of the Director of National Intelligence warning in its 2026 Annual Threat Assessment of the US Intelligence Community that China, Russia, Iran, and North Korea will continue to target the sector.

US director of national intelligence, Tulsi Gabbard, said nation state-backed threat groups typically target these systems to collect intelligence, create options for future disruption, and also for financial gain.

"China and Russia present the most persistent and active threats and are continuing their R&D efforts. North Korea’s cyber program is sophisticated and agile," she said.

"In 2025 alone, North Korea’s cryptocurrency heists probably stole $2 billion which is helping to fund the regime, including further development of its strategic weapons programs."

Recent targets have included Jaguar Land Rover (JLR), US water and wastewater systems and electrical subsystems, and the Ukraine power grid.

FOLLOW US ON SOCIAL MEDIA

Identity has become the central control plane for modern security. Indeed, according to Omdia, 75 percent of companies have increased their budgets around identity management in 2026, compared to 57 percent in 2025.

This increase is driven largely by agentic AI, as AI agents need identities and permissions to function. However, effective identity security now requires understanding how identity operates across the full environment, not within a single system.

The SANS Institute found that 68 percent of organizations can detect identity problems within 24 hours, but only 55 percent of companies can respond within that same timeframe. That gap matters because attackers operate inside that window, chaining access before remediation occurs. Understanding the risk is the starting point.

For partners, helping customers to spot gaps in their identity security is an opportunity. As environments become more fragmented and identity becomes the primary attack surface, how can you help your customers get ahead of these risks?

The future is hybrid

In the past, Microsoft Active Directory was the one place that controlled access and identity across the enterprise. That centralization made it a high-value target.

Today, identity is distributed. Developer identities and permissions live in GitHub. Endpoint platforms like Jamf manage device-linked access. Cloud identity providers such as Okta and JumpCloud extend identity across SaaS and infrastructure.

The result is more fragmentation across IT and the creation of trust relationships between systems. For example, a developer account in GitHub can use OpenID Connect to access cloud services and provision resources. In effect, one system is asserting identity into another. This is where risk begins to compound.

When an identity in one location has permissions in another, that identity becomes a dependency. Over time, identities accumulate permissions beyond their original purpose. For example, a developer account might have needed temporary deployment access but still retains full administrative rights. Alternatively, it may inherit access through group membership that exceeds what the individual account would otherwise receive.

In these environments, risk is rarely obvious when viewed system by system. The important question is not what an identity can access directly, but what it can reach through relationships. That is also exactly how attackers operate.

Responding to the identity challenge

Agentic AI has accelerated this problem. IDC forecasts that 40 percent of Global 2000 jobs will involve agentic AI by the end of 2026, while Gartner predicts that 40% of enterprise applications will embed task-specific AI agents by 2026, up from less than 5% in 2025. Gartner also predicts that 25 percent of these applications will experience multiple security issues annually.

Non-human identities (NHIs) have expanded rapidly alongside this shift. Managing them alongside human users requires an understanding of assigned permissions, how those permissions propagate, how they map to human operators, and how they evolve.

For partners, the starting point is establishing visibility into identities and their relationships. Looking at accounts and permissions provides a baseline. From there, the focus shifts to identifying which combinations of access create meaningful risk.

This involves prioritising remediation based on reachable impact, reducing misconfigurations, excessive privilege, and unintended access paths. This is not a one-time exercise with customers. Identity systems change continuously, especially with AI-driven automation. Static assessments degrade quickly in dynamic environments. This is particularly true for those non-human identities, where access is often not revisited.

Continuous visibility reduces both detection time and the gap between detection and remediation. Partners can also help customers protect critical assets by establishing priority zones where identity controls are more tightly enforced. This shifts security from reactive to structural. Coupled with regular reporting on identity changes, this creates a more effective and durable security model.

As enterprises expand AI and automation, identity becomes more critical. Just enumerating access is not sufficient to make systems more secure. Security depends on understanding how that access can be abused, as well as reducing the conditions that make abuse consequential.

Attackers do not compromise identities in isolation. They exploit relationships, using delegated rights, inherited roles, and cross-platform trust to move laterally and escalate privileges.

Experts have warned about the risks of paying ransoms after Instructure bowed to cyber criminal demands to avoid having stolen data published online.

The move comes after Canvas, a popular academic learning platform developed by Instructure, was breached by the ShinyHunters threat group last week.

More than 9,000 academic institutions across the US, UK, Canada, and Australia were impacted by the breach, which saw roughly 3.5 terabytes of data stolen by the ransomware group.

The move by Instructure marks the latest in a string of examples where organizations have chosen to play ball with hackers in the wake of a ransomware attack.

It’s a contentious topic for many in the security industry, and a tactic that is surprisingly common. Research from Absolute Security, published today, shows that more than half (57%) of CISOs would consider bowing to hacker demands to end a ransomware attack.

A key factor behind paying up, the study noted, lies in shortening potential downtime due to ransomware attacks. Nearly half (46%) ranked operational downtime as the most significant aspect of an attack.

To pay, or not to pay

The question of whether to pay up or not is a catch-22 for enterprises. Jeff Watkins of Leeds-based consultancy, NorthStar Intelligence, told ITPro that paying may appear to make sense for many given the potentially disastrous effects of data leaks.

“Paying cyber criminals may seem like a rational choice to avoid future data leaks, and in ransomware cases, where restoring from backups is not simple/feasible, it is often seen as necessary for business continuity,” he said.

Watkins pointed to the British Library attack in late 2023, which saw the institution refuse to pay a ransom. Hackers behind the attack subsequently released 500,000 files and recovery took several months – and at great cost.

Put simply, paying up often represents a small financial hit compared to the broader costs associated with recovery. Research last year, for example, found the average recovery cost for ransomware victims stood at $4.5 million.

But this tactic rests on trusting that the cyber criminals in question honor their side of the bargain, according to Watkins.

“There are risks involved in paying up, though,” he told ITPro. “There is that old adage, ‘there’s no honor amongst thieves’, and there is a risk that you simply lose your money, or they come back for more before deleting the data, providing a decryptor, or suppressing publication."

The Change Healthcare attack is a prime example of the risks involved with paying up, Watkins noted. The healthcare firm paid a $22 million ransom to the ALPHV/BlackCat group after a devastating 2024 attack, and they simply made off with the money.

RansomHub, an affiliate of ALPHV/BlackCat, still held data stolen in the breach and re-extorted the company.

Legal and moral ramifications

In addition to the operational considerations at play for enterprises, there are legal and moral ramifications.

The UK’s National Cyber Security Centre (NCSC) has been vocal in advising victims against paying up, while the US Cybersecurity and Infrastructure Security Agency (CISA) takes a similar stance.

In 2025, the UK government unveiled proposals aimed at banning ransom payments by public sector and critical national infrastructure (CNI) operators.

At the time, the government said the ban would “target the business model that fuels cyber criminals’ activities”.

Gary Barlet, public sector CTO at Illumio, said paying a ransom is often viewed as an “incentive for bad behavior” and simply places a bigger target on the back of those already affected by an attack.

“Cybersecurity professionals caution against this, because it signals to other threat actors that an organization is willing to pay if they can manage to steal data,” he told ITPro.

“Professionals worry that threat actors will then attempt to gain access to the same systems and demand even more in payments.”

Watkins echoed Barlet’s comments on threat actor incentivization, adding that choosing to pay effectively funds organized crime.

“This isn’t intended to be a criticism of the victims, as organizations pay because the choices are often ugly, not because they trust the criminals,” he said.

“They often face operational paralysis, patient/student/client harm, contractual penalties, regulatory exposure, reputational damage, and recovery costs far exceeding the ransom demand,” Watkins added.

“However, for as long as we allow organizations to pay ransom, the problems will escalate.”

Light on the horizon

There are positive signs that enterprise policies on ransomware attacks are changing, with many now refusing to play ball. As ITPro reported in August last year, research from Databarracks found just 17% of UK businesses paid ransoms in the wake of a breach.

This marked a steep decline compared to the year prior, in which more than a quarter (27%) opted to pay to recover stolen data. In 2023, nearly half (47%) chose to pay.

Enterprise backup strategies have helped on this front, the study noted, with victims choosing to recover instead of paying. More than half (57%) recovered data through backups after an attack across 2025.

Notably, Databarracks found enterprises are now three-times more likely to recover from backups than paying hackers, highlighting an increasingly tough approach.

FOLLOW US ON SOCIAL MEDIA

One-in-four organisations have exposed MySQL databases, according to new research, prompting calls for more robust developer security practices.

Intruder’s 2026 Attack Surface Management Index warned these databases are becoming an increasingly attractive target for threat actors, particularly ransomware groups.

Indeed, the study noted that 16% of Postgres databases are also dangerously exposed, alongside remote desktop (RDP) services, API documentation, and WordPress admin panels.

Attack surface exposures were categorized by HTTP panels, ports, services, databases, files and information facing the internet.

While exposed databases ranked as the leading attack surface issue, more than one-in-seven organizations reported exposed API documentation, ahead of RDP services - a common entry point for ransomware attacks.

Nearly half of organizations were found to have risky exposed ports and services, with RDP being the most commonly exposed. WordPress Admin (15%) and phpMyAdmin (8%) are also frequently left internet-facing, despite being intended for internal use only.

Notably, legacy services like SNMP (9%) and UPnP (8%) persist on the public internet, again despite being intended for internal networks.

Chris Wallis, CEO and founder of Intruder, said the findings should serve as a wake-up call for organizations engaging in risk data management security practices.

"Many of the exposures we examined don't even need a CVE to be exploited. For example, an exposed database or admin panel can be compromised through brute force or credential stuffing alone,” he said.

Database security in the spotlight

Intruder noted that lackluster data security practices come amid a perilous time for enterprises.

The study warned that the rise of autonomous AI models has slashed the time between vulnerability discovery and exploitation – and many organizations are struggling to keep up.

Midmarket organizations face the longest remediation times, averaging 56 days to close security gaps, making them nearly four-times slower than smaller enterprises.

There are stark differences between sectors, with banks remediating exposures in just 11 days and retail just ten, while insurance and pharmaceutical firms average more than 40 days.

With vulnerability exploitation expected to skyrocket due to the use of powerful new frontier AI models, Wallis said remediation windows are “open far too long”.

Security experts globally have issued repeated warnings on this front, particularly since the launch of Anthropic’s Claude Mythos model.

The company announced a gated release of the model to select industry partners in April amid fears the model could be used for nefarious purposes.

Wallis said the launch of Mythos has “fundamentally shifted” the cybersecurity landscape, meaning enterprises must now move faster than ever to curtail security risks.

“The security industry is seeing a major compression in the time between vulnerability discovery and exploitation,” he said.

“In this high-speed era, leaving a MySQL database or private API documentation exposed to the internet is an open invitation for automated, high-speed extortion.

FOLLOW US ON SOCIAL MEDIA

The UK government is urging organizations to sign a Cyber Resilience Pledge to help strengthen defenses against surging cybersecurity threats.

Set to launch later this year, the pledge will see businesses commit to making cybersecurity a board-level responsibility, sign up to the National Cyber Security Centre’s (NCSC) free Early Warning Service, and require Cyber Essentials certification across their supply chains.

The pledge is aimed primarily at medium and large organizations, although the government said it encourages firms of all sizes to take part.

"As threats evolve, businesses of all sizes need to step up and take practical action now," said cybersecurity minister Baroness Lloyd.

"The Cyber Resilience Pledge is a clear call for companies to strengthen their defences, protect their customers and play their part in keeping the UK secure and competitive."

Ministers have already written directly to some of the UK’s leading companies inviting them to sign up to the Cyber Resilience Pledge, and are urging other organizations to review the requirements and commit to it themselves.

Around £90 million in funding has been allocated by the government to help organisations conduct reviews.

Cyber Resilience Pledge comes at a critical time

The announcement comes alongside new government figures showing that the UK’s cybersecurity sector grew 11% last year to £14.7 billion.

There are 438 new cybersecurity firms, bringing the total to 2,603, up 20% from 2,165 last year. In terms of investment, £184 million was raised across 47 deals within dedicated cyber firms in 2025.

The sector now employs the equivalent of 69,600 full-time workers, up 3% since last year, in an increase of around 2,300 jobs; and the total gross value added (GVA) for the sector has reached around £9.1 billion, an increase of 17% since last year.

"Cybersecurity is now fundamental to economic growth, job creation and the resilience of the services people rely on every day," said Baroness Lloyd.

"The UK has a world‑class cyber sector that is creating skilled jobs and protecting our economy – and government is doing more by investing in its own defenses, legislating to require more of essential services and setting clear national standards."

AI is raising the stakes in cyber

The announcement also comes as the Cyber Security and Resilience Bill makes its passage through Parliament, following the King’s Speech, and forms part of an increasing government focus on cybersecurity.

The government has also set up the AI Security Institute which it said provides the most advanced capability of any government in the world for understanding frontier AI systems.

The Institute, for example, recently looked at frontier models like Claude Mythos and GPT 5.5, with ministers warning that traditional cyber protections alone are no longer enough.

As AI accelerates the pace and scale of cyber attacks, it said, organizations must now invest in smarter, more resilient systems that can limit the impact of breaches and keep ahead of attackers – rather than constantly reacting after the damage is done.

FOLLOW US ON SOCIAL MEDIA

The parent company behind academic management tool Canvas has issued an apology over disruption caused by a recent cyber attack, which impacted hundreds of institutions.

In a blog post on 11 May, Steve Daly, CEO of Instructure, said the company plans to introduce sweeping changes in the wake of the breach, insisting that Canvas is “fully operational and remains safe to use”.

Daly added that Instructure will continue providing assistance and guidance for institutions affected by the cyber attack.

“Rebuilding trust takes time,” he said. “We’re going to earn it back through consistent action and honest communication. We’re in this for you and your community.”

The apology comes after hundreds of schools and universities across the UK, Canada, Australia, US, and New Zealand were disrupted by a cyber attack waged by the ShinyHunters threat group.

The cloud-based academic management system is used by more than 8,000 institutions globally and has around 30 million active users.

Instructure first detected a breach on 1 May, but told customers it had taken steps to contain the incident.

In an advisory at the time, CISO Steve Proud warned data, including names, email addresses, student ID numbers, and messages between users had been impacted - which Daly confirmed in his recent blog post.

“This incident involved unauthorized access to part of our environment. The data fields involved include information like usernames, email addresses, course names, enrolment information and messages,” Daly wrote.

“Core learning data”, which includes course content, credentials, and student submissions, was not compromised in the breach, he added.

Canvas Cyber attack escalation

While Proud noted that the incident had largely been contained, the incident was compounded when ShinyHunters waged a follow-up attack, which saw user login portals defaced with a ransom note.

ShinyHunters claims to have gained access to around 3.65TB of Instructure data during the attack, which includes upwards of 275 million records from over 8,800 institutions.

Analysis of ShinyHunters activity ranks it as one of the most notorious ransomware groups in recent years. The group has claimed responsibility for a slew of attacks on major organizations such as Salesforce, Ticketmaster, and AT&T.

According to Daly, the Canvas attack saw ShinyHunters exploit a support ticket vulnerability in its Free for Teacher environment. The company has moved swiftly to contain the breach.

“We temporarily disabled Free for Teacher while we complete a full security review,” he said. “We know that’s disruptive, and we didn’t make that call lightly. But keeping the entire Canvas platform secure has to come first.”

“We didn’t deliver”

In his blog post, Daly said Instructure will continue providing updates and apologized for the company’s communication throughout.

“Over the past few days, many of you dealt with real disruption,” he wrote. Stress on your teams. Missed moments in the classroom. Questions you couldn’t get answered.”

“You deserved more consistent communication from us, and we didn’t deliver,” Daly added. “I’m sorry for that.”

The attack on Canvas comes during a busy period for academic institutions, with students in the midst of exams.

A slew of reports have detailed significant disruption for students on both sides of the Atlantic over the last week, with BBC coverage noting that Mississippi State University was forced to postpone exams.

As ITPro reported, students at the University of Oxford were unable to access papers and were forced to email lecturers for documents and results.

FOLLOW US ON SOCIAL MEDIA

Cyber criminals have been observed using AI to build a working zero-day exploit in a case Google researchers say is the first of its kind.

According to new research from Google Threat Intelligence Group (GTIG), a threat actor planned to deploy the zero-day in a “mass exploitation event” but was thwarted.

The company said that its “proactive counter discovery” may have prevented its use.

The zero-day in question bears all the hallmarks of an AI-generated exploit, researchers noted, largely due to the fact the script contained an “abundance of educational docstrings" as well as a hallucinated CVSS score.

Other tell-tale signs, such as a “textbook” Python format which is characteristic of LLM training data, also gave the game away for the threat actor.

Google noted that it does not believe its Gemini model was used to develop the zero-day exploit, but said it has “high confidence” another AI model was used.

John Hultquist, chief analyst at GTIG, said the discovery marks a significant moment in the use of AI for nefarious purposes.

“There’s a misconception that the AI vulnerability race is imminent,” he said. “The reality is that it’s already begun.”

Under the hood of the zero-day

According to GTIG, the vulnerability is classified as a two-factor authentication (2FA) bypass. However, researchers noted it requires valid user credentials.

What made the flaw stand out is that it stems from a “high-level semantic logic flaw where the developer hardcoded a trust assumption”.

Simply put, this particular flaw is the type that could be easily identified by large language models, and researchers deduced that the model’s reasoning capabilities allowed it to read the developer’s intent during development.

“While fuzzers and static analysis tools are optimized to detect sinks and crashes, frontier LLMs excel at identifying these types of high-level flaws and hardcoded static anomalies,” researchers noted.

“Though frontier LLMs struggle to navigate complex enterprise authorization logic, they have an increasing ability to perform contextual reasoning, effectively reading the developer's intent to correlate the 2FA enforcement logic with the contradictions of its hardcoded exceptions.”

Nation-state activity accelerates

The discovery by GTIG highlights the growing appeal of AI-based tools for cyber criminal groups and nation state-backed threat actors.

“Threat actors are leveraging AI to augment various phases of the attack lifecycle,” researchers said.

“This includes supporting the development of vulnerability exploits and malware, facilitating autonomous execution of commands, enabling more targeted and well-researched reconnaissance, and improving the efficacy of social engineering and information operations.”

Indeed, the Google threat intelligence wing said groups in China, North Korea, and Russia in particular are flocking to AI for vulnerability research and exploit development.

“These actors have leveraged sophisticated approaches toward AI-augmented vulnerability discovery and exploitation, beginning with persona-driven jailbreaking attempts and the integration of specialized, high-fidelity security datasets to augment their vulnerability discovery and exploitation workflows,” researchers said.

One threat group, tracked as UNC2814, was observed using expert persona prompting in Gemini to research remote code execution flaws in TP-Link router firmware and Odette File Transfer Protocol (OFTP) implementations.

Another group, APT45, was observed sending thousands of repetitive prompts to analyze different CVEs and validate PoC exploits, according to researchers.

This, GTIG added, is helping them to create a “more robust arsenal of exploit capabilities” that would be difficult to tackle for defenders.

FOLLOW US ON SOCIAL MEDIA

An update to Cyber Essentials, the UK government-backed cybersecurity certification scheme, came into place on April 27, but many businesses are still failing to take advantage of the UK government’s basic cybersecurity standards.

According to the government’s latest Cyber Security Breaches Survey, 43% of businesses reported having experienced a breach or attack in the last 12 months. Yet awareness of Cyber Essentials remains low, with just 12% of companies citing they knew about the standards.

It is with this in mind that the government launched a campaign in February, which urged companies to “lock the door” on cybercriminals by engaging with the scheme.

The Cyber Essentials annual update arrives in April each year to tighten up the standards so businesses are robust enough in the face of modern cyber threats. While the five core controls remain the same, experts say the 2026 update is one of the most significant changes to the scheme in years. As it comes into force, what do businesses need to know?

Significant changes

April’s update to Cyber Essentials was driven by the need to eliminate ambiguity around some requirements. The Cyber Essentials update focuses on “the actual core attacks”, such as credential theft, cloud account compromise and delayed patching – which are “consistently among the top entry points for attackers”, explains Aaron Bishop, CEO and founder of Novous.

“These are all areas where the previous revision of the requirements left too much room for interpretation,” according to Bishop.

Among the headline changes is a reduction in the patching window that mandates high-risk and critical security updates must be applied within 14 days of release.

For many firms, patching processes that have historically been tolerated as “good enough” will now be out of alignment with the standard, according to Jon Bance, chief operating officer at Leading Resolutions.

Following the update to Cyber Essentials, businesses, especially those with smaller IT teams, will need to be “much more deliberate” about how they track, prioritize and evidence patching activity across endpoints, servers and cloud services, he warns.

Another major update to Cyber Essentials will ensure firms are implementing multi-factor authentication (MFA) as a mandatory control for all cloud services. “If your cloud services offer two-factor authentication (2FA), MFA or single sign on (SSO), you must have it enabled – even if it is via SMS,” says Bishop,

The benefits of Cyber Essentials

In January this year, UK Digital Minister Liz Lloyd said that Cyber Essentials certified organizations are 92% less likely to make a claim on their cyber insurance than those without. Comments such as this show the business case for Cyber Essentials is becoming “more compelling” and difficult to ignore, says Bishop.

The updated Cyber Essentials certification helps reduce the risk of breaches by ensuring that key security measures are in place. In addition, organizations that achieve certification may be eligible for free cyber insurance, offering “an extra layer of financial protection in the event of an incident”, says Harry Mason, head of client services at Mason Infotech.

The scheme is becoming a supply chain prerequisite, with many larger organizations and public sector bodies requiring it from their suppliers and third-party contractors as a condition of contract. “So being certified can open your business to opportunities you otherwise wouldn't have been eligible for,” Bishop says.

It also provides a shared language between boards, technology teams and suppliers, helping conversations move away from “abstract cyber risk” to “practical, measurable controls”, according to Bance. For many, especially smaller firms, it therefore remains “one of the most cost effective ways of raising cyber hygiene”, he says.

Beyond compliance, the guidance provides “a highly practical, actionable security baseline”, agrees Ian Glennon, senior security solutions architect at Qualys. “Systematically applying these controls drastically reduces your attack surface and lowers your overall operational risk profile, which ultimately protects your bottom line.”

How businesses must respond to the Cyber Essentials update

With the update now in place, it’s a good idea to check where your business is on its Cyber Essentials journey. Adoption of Cyber Essentials is widespread, but maturity “varies considerably” across businesses, according to Bance.

Many organizations hold certification yet still operate reactively or informally, particularly around patching, asset visibility and administrative access, says Bance.

The update will expose that gap, he says. With this in mind, firms should avoid treating Cyber Essentials as “a one off annual activity”, Bance advises.

If you’re not doing it already, it is important to link Cyber Essentials into broader cyber or risk management practices. “This makes it easier to adapt with relatively minor changes,” Bance says.

Existing accounts have been given a six month grace period before needing to comply with the April changes. However, there is much to do, says Daryl Flack, partner at Avella Security.

Organizations should start by conducting a thorough audit of their existing Cyber Essentials scope covering all legal entities, in-scope devices, and out-of-scope justifications, according to Flack. He outlines the need to “implement mandatory MFA across every cloud service, whether free, paid or bundled”.

Meanwhile, companies should already have revised scoping declarations with board-level sign-off committing to ongoing compliance, while ensuring point-in-time assessments match certification dates. This should have been completed ahead of the April enforcement deadline, Flack advises.

In addition, with accelerated patching required, firms will need to ensure they’re installing high-risk and critical security updates or vulnerability fixes within 14 days, adds Flack.

When addressing the reduced patching window, the most important step is to review your current processes honestly, says Bance. “Particularly how quickly patches are applied and how evidence is captured.”

At the same time, Glennon points out that a fix does not exclusively mean deploying a vendor patch. “It includes any robust mitigation applied while a patch undergoes internal testing.”

Configuration changes, registry updates, disabling vulnerable services, or deploying specific scripts “all qualify as valid fixes”, he says.

Overall, to cope with the changes, it’s important that Cyber Essentials is treated as an “ongoing control set”, rather than a renewal exercise, according to Bance. “Those who embed it into day to day operations will find the update far less disruptive.”

Two US nationals have been sentenced to 18 months in prison for their part in running 'laptop farms' aimed at raising money for the North Korean government.

Matthew Issac Knoot, of Nashville, Tennessee, and Erick Ntekereze Prince, of New York, were sent company-issued laptops under stolen identities. They then installed unauthorized remote desktop software that allowed fake North Korean IT workers to appear as legitimate US-based employees.

The two worked separately, but according to the Department of Justice, between them generated more than $1.2 million in revenue for the DPRK and impacted nearly 70 companies in the US.

"These sentences hold accountable U.S nationals who enabled North Korea’s illicit efforts to infiltrate US networks and profit on the back of US companies,” said assistant attorney general for National Security John A. Eisenberg.

“These defendants helped North Korean ‘IT workers’ masquerade as legitimate employees, compromising US corporate networks and helping generate revenue for a heavily sanctioned and rogue regime. The National Security Division will continue to pursue those who, through deception and cyber-enabled fraud, threaten our national security.”

Prince helped at least three DPRK IT workers obtain remote employment at US companies between around June 2020 and August 2024. Prince used his company, Taggcar Inc, to supply 'certified' IT staff using false and stolen identities.

He also kept laptops provided by the victim companies at his New York home, installing remote access software without authorization to make it look as if the DPRK IT workers were working there.

Prince was sentenced to 18 months in prison, followed by three years of supervised release. He was also ordered to forfeit $89,000, the amount the DPRK IT workers paid him for his help.

Knoot, meanwhile, ran a laptop farm from his Nashville home between around July 2022 and August 2023, supplying North Korean IT workers to at least four US companies.

These firms paid the DPRK IT workers associated with Knoot’s laptop farm more than $250,000 for their work - most of which was falsely reported to the IRS and Social Security Administration under the name of the actual US citizen whose identity had been stolen.

He and his co-conspirators cost the victim companies more than $500,000 for auditing and fixing their devices, systems, and networks.

Fake North Korean IT workers are rampant

The crimes mark the latest in a continuing series of North Korean campaigns to supply fake workers and steal money for the regime.

Notably, these groups increasingly use voice-changing software during remote interviews to disguise their accents, or using the AI app Face Swap to place their faces in stolen identity documents and generate convincing headshots for CVs.

A host of organisations in the US have been affected by these campaigns over the last two years. As ITPro previously reported, cybersecurity company KnowBe4 unknowingly hired a fake IT worker, who immediately began loading malware as soon as they received their Mac workstation.

“This scheme shows how national security threats now enter through ordinary business systems. These defendants helped North Korean IT workers pose as legitimate employees, gain access to American companies, and generate money for a sanctioned regime,” said US attorney Jason A. Reding Quiñones for the Southern District of Florida.

"These were not paperwork violations. They were deliberate acts that exposed U.S. businesses, compromised trust, and supported one of the world’s most dangerous adversaries. These sentences send a clear message: if you help foreign actors infiltrate American companies for profit, you will face federal prison and lose the money you made.”

FOLLOW US ON SOCIAL MEDIA

Universities across the world are still experiencing difficulties after a cyber attack on the Canvas academic platform caused widespread disruption for staff and students.

Canvas is a cloud-based academic management system developed by Instructure, and is used by more than 8,000 institutions globally and around 30 million active users.

Staff and students at universities in the US, Canada, Australia, and the UK were severely disrupted when the platform was breached, with a ransom note allegedly from the ShinyHunters threat group appearing on login portals.

A host of UK-based academic institutions, including the Universities of Birmingham, Oxford, and Edinburgh were among those impacted in the breach.

Sources told ITPro that operations at the University of Birmingham are back online in the wake of the incident. However, the University of Oxford has warned students and staff that Canvas remains offline, with no confirmed date of return.

ITPro approached both institutions for confirmation, but did not receive a response by time of publication.

What happened with the Canvas cyber attack?

Instructure initially confirmed a breach occurred on 1 May, but had taken steps to contain and remediate the incident. According to the company, data exposed in the incident is believed to include “certain identifying information”, such as:

• Names
• Email addresses
• Student ID numbers
• Messages between users

Instructure’s chief information security officer (CISO), Steve Proud, said the company found “no evidence that passwords, dates of birth, government identifiers, or financial information were involved”.

On 2 May, Proud noted that the incident had been largely contained. However, ShinyHunters reportedly breached the company in a follow-up attack, defacing Canvas login portals at hundreds of institutions.

Analysis of the incident by Halcyon noted that ShinyHunters injected an HTML file that altered login screens, displaying a warning that the group will publish stolen data on 12 May if the company fails to pay a ransom.

On its leak site, ShinyHunters claims to have gained access to a sizable amount of company data – spanning 275 million records from 8,809 institutions, amounting to 3.65TB.

ShinyHunters ranks among one of the most prolific ransomware groups in recent years, having claimed responsibility for large-scale attacks on Salesforce customers, as well as AT&T and Ticketmaster.

Researchers at Halcyon noted that the group does not employ encryption during attack, but instead operates under a “pay or leak” extortion model.

“The group maintains a loosely decentralized structure with operational overlap among Scattered Spider (UNC3944), LAPSUS$, and Scattered LAPSUS$ Shiny Hunters (SLSH),” researchers said in a blog post detailing the incident.

ITPro has approached Instructure for comment.

Critical timing for ShinyHunters

The attack on Canvas comes at a critical time for institutions globally, with students preparing for exam season.

According to reports from BBC News, Mississippi State University was forced to postpone exams on Friday due to the incident. A meteorology student told the broadcaster that students were nearing exam deadlines when the platform was taken down.

The university has been engaging with students via email and told students it was affected by a “nationwide security incident”.

Sources told ITPro that students at the University of Oxford have been experiencing similar difficulties, with some unable to access papers and having to email lecturers for attached documents.

Universities in a host of other US states, as well as in Canada, New Zealand, and Australia have also experienced significant disruption.

The University of Sydney, for example, told students that Canvas was unavailable on Friday and warned students not to log in.

FOLLOW US ON SOCIAL MEDIA

A fake Claude AI website is spreading a trojanized 'Claude‑Pro' Windows installer that secretly distributes a newly-identified backdoor.

The domain mimics the official site for Anthropic’s Claude AI tool, and visitors who download the ZIP archive are sent a copy of Claude that appears to install and runs as expected.

However, researchers at Malwarebytes found it deploys a PlugX-like malware chain, dubbed Beagle, that gives attackers remote access to the system.

The ZIP contains an MSI installer that installs to a path designed to mimic a legitimate Anthropic installation, complete with a reference to Squirrel, the update framework that real Electron-based applications like Claude use.

A tell-tale giveaway for developers is that this contains a misspelling: ‘Cluade’.

While the legitimate application runs in the foreground, the VBScript quietly copies three files from the SquirrelTemp directory into the Windows Startup folder.

"This is a textbook DLL sideloading attack, a technique catalogued by MITRE as T1574.002. NOVUpdate.exe is a legitimately signed G DATA antivirus updater. When it executes, it attempts to load a library called avk.dll from its own directory," researchers explained..

"Normally, this would be a genuine G DATA component, but here the attacker has substituted a malicious version. Signed sideloading hosts like this can complicate detection because the parent executable may appear benign to endpoint security tools.

Victims are kept in the dark, because after deploying the payload files, the VBScript writes a small batch file called ~del.vbs.bat that waits two seconds, then deletes both the original VBScript and the batch file itself.

"This means the dropper is gone from disk by the time a user or analyst goes looking for it. The only artifacts that persist are the sideloading files in the Startup folder and the running NOVUpdate.exe process," Malwarebytes said.

"The script also wraps the entire malicious payload section in an On Error Resume Next statement, silently swallowing any errors so that failures in the deployment do not produce visible error dialogs that might alert the victim."

What is DLL sideloading?

DLL sideloading is a technique favored by PlugX, a malware family that Sophos has been tracking for 14 years.

As the firm points out, PlugX has multiple variants and has been associated with several threat actor groups, meaning that attribution isn't clear-cut.

On top of this, ShadowPad, another backdoor employing DLL sideloading, has a number of code overlaps with PlugX, to the extent that it could be considered an evolution of it.

"Most of the techniques described here are relatively well known and have been seen before, from spoofing a legitimate installer website to side loading using a signed executable. Interestingly enough what is unusual is that it also installs a working copy of Claude which is rather large," said Max Gannon, cyber intelligence team manager at Cofense.

"The installation and usage of a program that is resource intensive can also help to disguise other ongoing background activity. The use of a legitimate program, cleanup utilities, running in memory, and persistence mechanisms all indicate that the threat actors distributing this malware intend it for long term persistence and use."

FOLLOW US ON SOCIAL MEDIA