OpenAI releases final report on AI hacking incident

OpenAI has released a final report detailing a security incident in which the company deployed 700 AI agents to conduct simulated hacking attacks against its own systems. This experimental exercise was designed to test the resilience of its software against sophisticated cyber threats. By using these autonomous agents to identify vulnerabilities, the company aims to better understand how AI might be used both to secure and to compromise complex digital infrastructure.