Claude helped researchers access OpenAI employee accounts and code

Security researchers used Anthropic’s Claude to find and exploit weaknesses in OpenAI’s systems, taking over employee accounts and reaching an internal code repository before reporting the flaws. Claude was not itself the opening into OpenAI; it served as an assistant for the work of finding and linking vulnerabilities. Because the researchers disclosed the weaknesses after gaining access, the episode is a security demonstration rather than an account of a criminal attack. But it illustrates the growing concern: AI tools can help assemble sophisticated attacks faster, raising the stakes for companies to stop a small flaw from turning into broad internal access.