OpenAI Adds Violence Safeguards to ChatGPT and Plans Safety Report
OpenAI is tightening how ChatGPT handles violence-related requests, adding limits on content about past mass attacks and on weapons information for teens. The company also committed to a recurring Community Safety Report, with the first edition due in early 2027.
- OpenAI will publish a Community Safety Report, first in early 2027 covering the second half of 2026, then every six months.
- The company reports over 200 law enforcement referrals in Q3 2026.
- ChatGPT will give less detailed answers about past mass attacks for all users, regardless of age.
- ChatGPT will be able to automatically end chats that show fixation on violence, and repeated cases may lead to a temporary account pause.
- Teens face further limits on weapons and security system information, and parents are to be notified of pauses for repeated violence discussions.
A recurring report on community safety
According to OpenAI, the new Community Safety Report will cover violence prevention and other community safety issues, as well as how effective the company's safeguards are. It will explain how OpenAI trains its systems to detect, investigate and respond to such issues. The company says it will share anonymized examples of how violence-related harm appears, highlight new mitigations and policy changes, and include relevant data.
The first report is planned for early 2027 and will cover the second half of 2026. Updates will follow every six months. OpenAI also said it made over 200 law enforcement referrals in Q3 2026 for imminent threats of violence, and that it relies on mental health and behavioral experts to assess difficult cases.
Tighter limits for all users
OpenAI says its models are already trained not to give instructions that would help someone plan or carry out violence. It is now adding restrictions on how ChatGPT answers questions about past mass attacks and the people behind them. The model is directed to give less detailed information in those areas, even if it is available from other online sources.
The company says users should still be able to ask ChatGPT to help them understand such events, how victims and communities were affected, and how to recognize warning signs. The restrictions target reconstructions of past attacks, including details about weapons and equipment and how they were obtained, and detailed comparisons of scale of harm that could amplify an attacker's notoriety. According to OpenAI, these limits apply regardless of a user's age or reason for asking.
OpenAI is also building on a May update in which ChatGPT used short, factual summaries of earlier warning signs to respond more carefully in later chats. When recent chats show signs of possible violence, ChatGPT will respond more conservatively when sharing information that could be misused. The company is working to apply those stricter limits when it predicts someone is under 18, even without signs of violent intent.
OpenAI is also working with mental health and violence prevention experts on automated interventions. If a conversation shows signs that someone is fixating on violence or on seriously harming another person, ChatGPT will be able to automatically end the chat. If multiple chats are ended for that reason, the account may be temporarily paused. Accounts that violate the violence policies are banned.
More protections for teens
ChatGPT for Teens applies age-appropriate safeguards for users under 18. OpenAI is adding limits on weapons-related information teens can receive, even when a question appears harmless. The company says it is extending existing restrictions to practical information about weapons and security systems that could be misused to facilitate violence, even when there is no sign of violent intent. Teens can still use ChatGPT to help prevent and de-escalate violence, seek support and connect with a trusted adult.
Age prediction, launched last year, estimates whether a user may be under or over 18 from signals such as behavior patterns. OpenAI says it has strengthened the system to keep assessing new signals and move teens into the under 18 experience when appropriate. Users placed there incorrectly can confirm they are adults through Persona, OpenAI's age verification partner, using what the company describes as a fast, privacy-protective process.
Parental controls let parents link accounts with their teens and manage safety settings. Linked accounts can receive safety alerts when specially trained reviewers identify a serious self-harm concern or behaviors that may be associated with disordered eating. In July, those alerts were expanded to notify parents if a teen's account was deactivated for violating violence-related usage policies. OpenAI says it is working on notifications for parents when a teen's account is paused for repeatedly discussing violence.
What happens next
OpenAI says the safeguards described will begin rolling out in the next few weeks and will continue to be refined. The company describes the work as ongoing, noting that recognizing warning signs reliably and deciding when and how to intervene remain open challenges. For users, the practical changes are more conservative answers on violence-related topics, the possibility of ended chats or paused accounts in cases of fixation on violence, and stricter weapons-related limits for teens, with the planned report meant to show how well these measures work.
OpenAI: Community safety report and new protections for teens
