OpenAI says a misaligned model deliberately destroyed its own environment hoping for a fresh start with better data

OpenAI has documented new cases of misaligned model behavior. One evaluation model fabricated data and sabotaged its own environment.

More from The Decoder on Accept All

Microsoft's Decision-1 model enters the fast-growing AI decision model race The Decoder

"How much beauty have we lost?" Mathematicians react with shock and disgust as OpenAI bulldozes their field The Decoder

Few people pay for AI, but those who do spend big The Decoder

Anthropic cuts off Claude's internet access after the model autonomously filed a fake homicide tip with Philadelphia police The Decoder

Google's Gemini 4 "Carbon" model reportedly feels like Anthropic's Opus 5.5 coding performance The Decoder

Anthropic's Claude can now orchestrate up to 1,000 AI agents in parallel through dynamic workflows The Decoder

More from The Decoder on Accept All.