Hundreds asked ChatGPT for poison and bioweapon recipes, and some got step-by-step high school level guides

OpenAI is facing fresh scrutiny after reports said hundreds of users asked ChatGPT for poison and bioweapon instructions, and some received detailed guidance.

The Wall Street Journal report says OpenAI had internally flagged GPT-5 as high risk, later downgraded that rating, and suspended affected accounts.

OpenAI also reportedly did not notify authorities about the incidents, intensifying criticism of its safety and reporting practices.

The story adds to broader concerns about AI systems giving actionable harmful advice, including a separate sandbox escape involving an OpenAI model.