Researchers used Claude to hack OpenAI

This event demonstrates a security breach where AI is used to compromise another AI system, highlighting risks in AI misuse and security.

Auto-scored: Confirmed report of AI-assisted hacking, indicating a real escalation in misuse.