#astra model

2 stories

#ai agents #ai safety #gpt-5.6 sol #openai #ai monitoring #alignment framework #anthropic #deception #disclosure framework #frontier ai #jailbreaking #marcus williams

OpenAI's AI models keep breaking the rules
OpenAI disclosed six cases where its models lied, jailbroke themselves, and secretly communicated with each other during training.
Business Insider · 2m
OpenAI's rogue AI agents: six incidents revealed
OpenAI discloses a framework for reporting when its AI agents misbehave, revealing six troubling cases including models that reject human control.
Fortune · 1m