OpenAI Will Publish a Twice-Yearly Safety Report and Tighten ChatGPT Limits on Violence and Weapons
ChatGPT will stop detailed reconstructions of past mass attacks for all users and restrict practical weapons information for teens, with rollouts starting in the coming weeks.
Published
Last updated / Fact-checked on
OpenAI said on October 30, 2026, that it will publish a Community Safety Report every six months, starting in early 2027 with a report on the second half of 2026. It also announced new ChatGPT limits: no detailed reconstructions of past mass attacks for any user, and tighter restrictions on weapons-related information for teens. The safeguards begin rolling out over the next few weeks.
Image: Illustrative photo of a black smartphone held in a hand. Dennis Cortés via Unsplash / Wikimedia Commons — CC0 1.0
A new transparency report
OpenAI said the report will cover violence prevention, other community safety issues and how well its safeguards work. It will explain how the company trains its systems to detect, investigate and respond to these issues. It will also include anonymized examples of how violence-related harm shows up, new mitigations and policy changes, and supporting data. (The announcement also refers to it once as a "Community Safeguards Report.")
The company said it made more than 200 referrals to law enforcement in the third quarter of 2026. It said these referrals are helping disrupt violent plots.
Limits for every user
OpenAI's models are already trained not to help anyone plan or carry out violence. The company is now restricting how ChatGPT discusses past mass attacks and the people who carried them out, even when the information is available elsewhere online. These limits apply regardless of a user's age or reason for asking.
| ChatGPT will still help with | ChatGPT will restrict |
|---|---|
| Understanding what happened | Reconstructions of past mass attacks |
| How victims and communities were affected | Weapons and equipment used, and how they were obtained |
| Recognizing warning signs and preventing violence | Detailed comparisons of the scale of harm that could amplify an attacker's notoriety |
OpenAI is also adding responses to warning signs that build up over several conversations. In May, it described how ChatGPT can use short, factual summaries of earlier warning signs. Now, when recent chats show signs of possible violence, ChatGPT will respond more conservatively with information that could be misused.
When a conversation shows someone fixating on violence or on seriously harming another person, ChatGPT will be able to end the chat automatically. If that happens repeatedly, OpenAI may temporarily pause the account. Accounts found to violate its violent-activity policies are banned.
Stricter rules for teens
ChatGPT for Teens already applies age-appropriate safeguards for users under 18. OpenAI will now restrict practical information about weapons and security systems that could be misused, even when a question seems harmless and there's no sign of violent intent. Teens can still use ChatGPT to prevent and de-escalate violence, get support and connect with a trusted adult.
OpenAI said it is also working to apply the more conservative warning-sign limits whenever it predicts a user is under 18, even with no signs of violent intent.
Other parts of the teen system:
- Age prediction. Launched last year, it estimates whether a user is under 18 from signals such as behavior patterns. OpenAI says it now reassesses as people use ChatGPT.
- Adult verification. Adults wrongly placed in the teen experience can confirm their age through Persona, OpenAI's verification partner, to remove the extra protections.
- Parent alerts. Linked parent accounts may receive alerts about serious self-harm concerns or possible disordered eating. In July, OpenAI added alerts when a teen's account is deactivated for violating its violence policies. Notifications for violence-related account pauses are in development.
Why it matters
The changes affect one of the most widely used AI products. Adults researching past attacks may get less detail about weapons, equipment and scale. Teens will hit firmer limits on weapons and security-system questions, even routine ones. The report could give outside observers a regular look at how OpenAI's safeguards perform, though its value will depend on what data the company chooses to publish.
What to watch:
- Early 2027: the first report, covering July through December 2026
- The coming weeks: the rollout of the new limits
- Parent notifications for paused teen accounts
More coverage is in our Artificial Intelligence hub.
How we reported this
We checked every claim against OpenAI's announcement and corrected two features the draft described as already in place; OpenAI says it is still working on them. We removed the draft's appended summary and tips sections and condensed repeated company statements.
Published and fact-checked October 30, 2026.
Sources
DailyMech used AI tools to help research and draft this article; it was fact-checked by our editors.