OpenAI adds teen safeguards, violence limits and safety reporting

OpenAI said it is expanding safety controls in ChatGPT, including tighter limits on violence-related content, additional protections for users under 18 and a planned public Community Safety Report. The company said the changes are aimed at reducing misuse while improving transparency about how its systems detect and respond to harmful activity.

Evidence and context

In its Oct. 30, 2026 announcement, OpenAI said it made more than 200 law-enforcement referrals in Q3 2026 and described those referrals as part of its approach to addressing imminent threats of violence. The company said it relies on mental health and behavioral experts to help assess difficult cases and that its systems are designed to detect, review and escalate credible threats.

OpenAI said it will launch a Community Safety Report focused on violence prevention and other community safety issues, as well as the effectiveness of its safeguards. The first report is scheduled for early 2027 and will cover the second half of 2026, with updates planned every six months afterward. The company said the report will include anonymized examples, policy changes, new mitigations and supporting data.

The announcement also tightens how ChatGPT responds to questions about past mass attacks. OpenAI said the system will provide less detailed information in those areas, including details about weapons and equipment used and how they were obtained, as well as detailed comparisons of harm that could increase an attacker’s notoriety. The company said these limits apply regardless of age or intent.

For teens, OpenAI said it is adding further restrictions on weapons-related information, including practical information about weapons and security systems that could be misused to facilitate violence. The company said teens can still use ChatGPT for prevention, de-escalation, support and reaching a trusted adult. OpenAI also said it has strengthened age prediction so it can continually assess new signals and move users into the under-18 experience when appropriate, with a privacy-protective verification path for adults who are mistakenly placed there.

OpenAI said it is also building on earlier work that uses short factual summaries of warning signs from prior conversations, and that it may automatically end chats when a conversation shows signs of fixation on violence or serious harm. If that happens repeatedly, the company said it may temporarily pause the account. It added that parents linked through its parental controls may receive safety alerts, including notices in some cases of deactivation for violence-policy violations, and that it is working on notifications if a teen’s account is paused for repeated discussion of violence.

The company said the safeguards will begin rolling out in the next few weeks. It did not provide a full deployment schedule for every feature, and some measures, including automatic chat interruption and possible account pauses, were described as ongoing work rather than universal features already in place.