On October 30, 2026, Reuters reported that OpenAI had announced new ChatGPT safeguards for teens, including limits on practical information about weapons and security systems that could be misused. The restriction was described as applying even when a teen’s question showed no apparent violent intent. OpenAI told Reuters the rollout would begin in the next few weeks after the announcement.
Which ChatGPT safeguards were announced for each audience?
The reported weapons-information restriction is for teens. Reuters described separate violence-related measures as applying to all ChatGPT users:
| Audience | Announced measure |
| Teens | Restrict practical information about weapons and security systems that could be misused, even when a question shows no apparent violent intent. |
| All users | ChatGPT may end chats that appear to fixate on violence; after multiple such chats, accounts may be temporarily paused. Responses will also give less detail about past mass attacks, including weapons used and how they were obtained. |
The reported teen restriction targets practical information, rather than every question that mentions a weapon. The measures for all users address conversations that appear to fixate on violence and details about past attacks.
What age verification changes for adults
OpenAI says eligible accounts identified as belonging to someone under 18 can be moved into ChatGPT for Teens. An adult whose account is placed in the teen experience can verify their age through Persona. If Persona confirms the account holder is at least 18, OpenAI says it removes teen protections; the change may take a short time. OpenAI’s age-prediction guidance does not identify whether the newly announced weapons restriction is among the protections removed.
Depending on the country, Persona may ask for a live selfie, government ID, or both. OpenAI says Persona deletes an uploaded ID or selfie within seven days, and that OpenAI receives neither image. It receives age-related information used to select the account experience.
What parents can see in teen safety alerts
OpenAI says optional alerts for linked teen accounts may be sent when its systems detect acute self-harm or violence-related risk. The alerts are reviewed by trained experts, and OpenAI says parents do not receive the teen’s prompts or generated content. OpenAI’s teen safety page describes those alerts; Reuters separately reported that parents would be notified if an account were paused after repeated violence-related chats, without specifying what that notification would contain.
