OpenAI tightens ChatGPT limits on weapons and mass-attack details, and will publish a twice-yearly community safety report

OpenAI announced on Friday 30 October new restrictions on how ChatGPT handles violence-related questions, extra limits for users under 18, and a regular public report on its safety work.

The company said ChatGPT will now give less detailed information about past mass attacks and the people who carried them out, even when that information is available elsewhere online. It said it was restricting reconstructions of such attacks, including details of the weapons and equipment used and how they were obtained, and comparisons of the scale of harm. OpenAI said these limits apply “regardless of a user’s age or reason for asking”, while people can still ask ChatGPT to help them understand such events and their impact on victims.

When recent chats show warning signs of possible violence, ChatGPT will answer more conservatively, OpenAI said. It said ChatGPT will also be able to end a conversation automatically when it shows signs that someone is fixating on violence or on seriously harming another person, and that accounts where this happens repeatedly may be temporarily paused.

For teens, OpenAI said it is restricting practical information about weapons and security systems that could be misused, even when there is no sign of violent intent. It said its age-prediction system now keeps checking new signals and moves users it predicts are under 18 into the teen experience; adults placed there by mistake can confirm their age through its verification partner, Persona. OpenAI said it is also working on notifying parents if a linked teen account is paused for repeatedly discussing violence.

OpenAI said it made more than 200 referrals to law enforcement in the third quarter of 2026. Its first Community Safety Report, covering the second half of 2026, is due in early 2027, with updates every six months. The company said the new safeguards will start rolling out “in the next few weeks”.