OpenAI Expands ChatGPT Safeguards for Teens and Past Mass Attacks

OpenAI will limit ChatGPT’s practical guidance on weapons and security systems that could facilitate violence for users under 18, while barring all users from reconstructing details of past mass attacks. The changes expand the chatbot’s safety rules for teens and add restrictions that apply regardless of a user’s age.

The teen limits cover practical information about weapons and security systems when that guidance could be misused to enable violence. A request for that material may be restricted even if it appears harmless; the rule does not require signs of violent intent in the conversation. The changes build on existing rules against helping users plan or carry out violence and limits on helping teens find or obtain weapons.

OpenAI’s age-estimation system will continue evaluating signals during use and may identify or estimate that someone is under 18. Adults mistakenly placed in the teen experience can verify their age through Persona to remove the additional protections. OpenAI announced the expansion Oct. 30 and plans to start rolling it out over the next several weeks, extending its earlier limits on ChatGPT’s weapon answers for teens.

Separate safeguards will apply to everyone, regardless of age or reason for asking. ChatGPT will limit attempts to recreate past mass attacks, including the weapons and equipment involved, how they were obtained, and detailed casualty comparisons that could amplify attackers’ notoriety. People can still use the chatbot to understand effects on victims and communities, identify warning signs and learn about preventing violence.

ChatGPT may automatically end a conversation when it detects fixation on violence or intent to seriously harm someone. Repeated endings associated with the same account may lead to suspension. OpenAI also plans to notify parents when a teen account is suspended after repeated discussions of violence. Teens can still use ChatGPT to seek support, help prevent or de-escalate violence, and reach a trusted adult.

OpenAI referred more than 200 cases involving imminent threats of violence to law enforcement in the third quarter of 2026. The company plans to release its first community safety report in early 2027, covering the second half of 2026, and update the report every six months.