AI news · OpenAI · child safety
Common Sense Media rates ChatGPT for Teens an unacceptable risk
The nonprofit says parent alerts stayed silent in self-harm chats; OpenAI disputes how the research was run.
Common Sense Media rated ChatGPT for Teens an “unacceptable risk” on Wednesday 7 October 2026, the BBC reports. OpenAI launched the teen version in August with parental alerts, limits on emotional dependence and other safeguards. The nonprofit says several of them fail, and that under-18s should be kept off ChatGPT until they work.
Its researchers held hour-long conversations about suicidal thoughts, self-harm and eating disorders on new, parent-linked accounts and say no alert reached a parent. Only older accounts with weeks of history triggered one. ChatGPT refused sexual roleplay, but it still did schoolwork and talked like a friend, the BBC reports.
TechCrunch focuses on engagement. The bot sent teens to a trusted adult in 94% of crisis prompts about other people, but rarely when the risk was the teen’s attachment to ChatGPT itself. Across nearly 2,000 prompts, testers saw two break reminders.
OpenAI says much of the testing may have begun and concluded before parental controls were fully active. Its own data, released the same day, puts average teen use under 15 minutes a day.
For anyone building AI products for young users, the dispute shows outside groups now grading safety features against the maker’s own promises.
Sources
- BBC News: OpenAI says teen ChatGPT use limited but research finds it an ‘unacceptable risk’. Checked
- TechCrunch: ChatGPT for Teens keeps teens talking, even during mental health crises. Checked
Alexandre S.
Alexandre S. is a creative ad strategist and copywriter with ten years' experience in digital marketing across multiple channels.