AI's operational blind spots

Businesses continue to embrace generative AI at a rapid pace. From copilots and chatbots to autonomous agents, an exciting world of new technology is opening up.

The race to adopt these AI tools isn't without its perils, though. The mantra "garbage in, garbage out" is even more poignant when these systems come into play, with the potential for a cycle of bad data presenting a real risk to organizations.

On this week's episode, Tim Pfaelzer, SVP and GM EMEA at Veeam, discusses operational blind spots, what they mean for businesses, and how they can balance their desire for a technical edge with data governance.

Highlights

"Now we're in the third era, which is agentic, which is kind of hyper-accelerating everything that's there to touch your data. It's not like somebody doing something to attack your data. It is multiple agents running simultaneously across hundreds of apps and all the different systems to attack your data within day number one."

"There are tons of examples that you would probably call worst-case scenario. If you think about, you know a large online retailer that lost all of its backups, all of its production data because of a hallucinating AI agent within nine seconds, resulting in about 10 million lost orders and a 72-hour outage. Those are the type of things we're talking about. And now, if you ask about the worst case, it's not only the direct orders that you lost. It's not only the direct downtime that you have, but there's a next currency to it, and that next currency is the currency of trust."

Links

• AI-assisted software development means security teams need an 'engineering-first' mindset
• IT leaders are facing major work device blind spots – and it's putting security at risk