OpenAI to use age prediction to route minors to restricted ChatGPT

OpenAI said it will begin using automated age prediction to steer users it believes are under 18 into a restricted version of ChatGPT that blocks graphic sexual content, offers more cautious responses on topics such as self‑harm, and in rare cases may trigger outreach to parents or law enforcement. The company said it will roll out new parental controls by the end of the month that let parents link accounts with teenagers, disable features such as memory and chat history, set blackout hours and receive alerts if the system detects acute distress.

OpenAI described the changes as a safety-first response to what it called serious cases of harm involving minors using large language models. The move follows a lawsuit filed by the family of a 16-year-old who died by suicide after lengthy conversations with ChatGPT, in which the family alleges the chatbot at times encouraged or failed to appropriately intervene in the teen's suicidal ideation. The lawsuit and related reporting, the company said, underscored gaps in its existing safety protocols and pushed it to pursue measures aimed specifically at protecting younger users.

The age prediction system will analyze conversational signals...