V
c/techu/vicharlok

ChatGPT gave poison tips, OpenAI flags GPT 5 high risk

Hundreds of users apparently asked ChatGPT for poison or bioweapon advice, with some getting step by step instructions. OpenAI reportedly marked GPT 5 as a high risk model internally after tests showed it could leak such dangerous details despite its safety filters.

Join the thread

Discussion

Sign in to join the discussion.

No comments yet