OpenAI to pay up to $20k in rewards through new bug bounty program
The move follows a period of unrest over data security concerns
OpenAI has unveiled a new bug bounty program offering rewards for security researchers if they can uncover vulnerabilities in its products.
In an announcement on Tuesday, the California-based AI firm said the bug bounty scheme is “essential to our commitment to develop safe and advanced AI” and deliver services that are secure, reliable, and trustworthy.
As part of the initiative, OpenAI said it will offer a tiered reward system based on the severity of bugs uncovered by researchers.
Rewards can range from as little as $200 for low-severity flaws with a maximum reward of $20,000 for “exceptional discoveries”.
“The OpenAI Bug Bounty Program is a way for us to recognize and reward the valuable insights of security researchers who contribute to keeping our technology and company secure,” the firm said in a statement.
“We invite you to report vulnerabilities, bugs, or security flaws you discover in our systems. By sharing your findings, you will play a crucial role in making our technology safer for everyone.”
Researchers participating in the new initiative will be able to disclose vulnerabilities or flaws through a partner organisation, Bugcrowd.
Bugcrowd will manage the submission and reward process, which OpenAI said is designed to “ensure a streamlined experience for all participants”.
ChatGPT vulnerability concerns
The move from OpenAI follows a period of unrest over security-related issues at the generative AI firm, which has close ties with Microsoft.
Last month, the company revealed that a bug in ChatGPT led to a leak of users' data.
RELATED RESOURCE
SOC modernisation and the role of XDR
How to cope with increasing threats and IT sprawl
This flaw meant that ChatGPT Plus users began seeing user email addresses, subscriber names, payment addresses, and limited credit card information.
The issue prompted the company to temporarily take the chatbot offline to work on a fix.
“The bug was discovered in the Redis client open-source library, redis-py,” OpenAI explained in a post at the time.
“As soon as we identified the bug, we reached out to the Redis maintainers with a patch to resolve the issue.”
-
An ‘unprecedented cyber incident’: How OpenAI models breached Hugging Face – and why it could herald a ‘new phase of AI-powered cyber crime’News The incident should serve as a stark warning on the dangers of AI agents, according to cyber experts
-
The agents you use to beef up cybersecurity could be turned against you – ‘Friendly Fire’ attacks can manipulate OpenAI and Anthropic models into running malicious codeNews Research shows agents can be fooled into executing malicious code while performing security reviews of third-party software
-
OpenAI expands 'Daybreak' cyber program: New tools, partnerships, and a cyber-focused GPT-5.5 aim to help 'patch the world'News The company has added new tools, signed up partners, and released its GPT-5.5-Cyber model more widely
-
Hackers are capitalizing on AI hype to ramp up social engineering attacks – and they're using big brands like Anthropic, OpenAI, and DeepSeek as ‘bait’ to lure victimsNews Microsoft says cyber criminals are impersonating popular AI platforms to deliver malware
-
AI is shrinking attack windows, and it’s forcing a complete rethink of cyber resilience – here’s how organizations can prepareNews Commvault has urged companies to improve their business continuity and resilience plans in the face of flaws spotted by AI
-
Everything you need to know about ChatGPT’s new Advanced Account Security featuresNews OpenAI has introduced new tools to tightening up access to ChatGPT, Codex, and its other AI tools
-
Anthropic targets vulnerability detection gains with Claude Security public beta — here's what users can expectNews The Claude Mythos developer is aiming for a more limited approach to cyber tooling for public consumption
-
Researchers warn millions of RDP and VNC servers are wide open to exploitationNews Researchers at Forescout spotted millions of RDP and VNC servers exposed online
