OpenAI commits to a biannual community safety report and tightens weapons limits for teens, referring over 200 cases in Q3
OpenAI announced on October 30 two community safety moves, a community safety report and tighter restrictions for minors. It said it referred more than 200 cases to law enforcement in the third quarter of 2026 to intervene in plots of violence.
OpenAI announced on October 30 two community safety moves, a community safety report and tighter restrictions for minors. It said it referred more than 200 cases to law enforcement in the third quarter of 2026 to intervene in plots of violence. The report will cover violence prevention and other community safety issues and the real effectiveness of safeguards, explaining how systems are trained to detect, investigate and respond to community safety risks, sharing anonymized examples of how harmful behavior manifests, highlighting new mitigations and policy changes, and including supporting data. The first report will be published in early 2027 covering the second half of 2026 and updated every six months thereafter. Beyond not training models to give instructions for planning or carrying out violence, OpenAI tightened how ChatGPT responds to information about past mass attacks and their perpetrators, providing fewer details even when such information is available elsewhere, restricting details about how weapons and equipment were obtained and comparisons of casualty scale that could amplify an attacker's notoriety, regardless of user age or reason for asking. When recent chats show warning signs of possible violence ChatGPT responds more conservatively, and stricter limits apply early once the system predicts a user is under 18. If a conversation shows fixation on violence or seriously harming someone, ChatGPT can end the chat and accounts repeatedly ended for that reason may be temporarily suspended. For users under 18, limits extend to practical weapons information including security systems that could facilitate violence, while teens can still use the service to de-escalate conflicts and seek support. OpenAI said it strengthened age prediction and that users wrongly placed in the teen experience can confirm adulthood through Persona to have protections removed.