OpenAI says ChatGPT will end chats fixated on violence and tighten limits for teens
The company says it referred more than 200 cases to law enforcement in the third quarter and will publish safety reports twice a year.
OpenAI said Oct. 30 that ChatGPT will soon be able to end, on its own, conversations in which a user keeps focusing on violence or on seriously harming someone. OpenAI said the goal is to interrupt that pattern in the moment. If ChatGPT ends several chats, the company may temporarily pause the account.
OpenAI did not say what message or warning users will see when a chat is ended. It said it is working with mental health and violence prevention experts on the changes. Since July, parents with linked teen accounts have been alerted if the teen's account is shut down for violence, and OpenAI said it is working on alerts for pauses too.
OpenAI said it made more than 200 referrals to law enforcement in the third quarter of 2026. It plans to publish a safety report every six months, starting in early 2027.
For all users, ChatGPT will limit details about past mass attacks, such as which weapons were used. Teen users will face limits on practical information about weapons and security systems, even without signs of violent intent.
ChatGPT guesses whether a user is under 18 from signals such as how they use it. Adults wrongly placed in the teen version can prove their age through Persona, an outside company that checks identities and ages for online businesses, often with a photo ID or a selfie. Once confirmed, the teen limits come off.
Source: OpenAI, "Community safety report and new protections for teens," Oct. 30, 2026; Persona, age assurance.