Overview
- Common Sense Media’s Youth AI Safety Institute published its assessment on Oct. 7 and rated ChatGPT for Teens an “Unacceptable Risk,” urging OpenAI to stop marketing the feature and to block teen access until independent testing verifies fixes.
- The institute’s tests found parental notifications often failed when testers discussed suicide, self‑harm or eating disorders, with linked parent accounts receiving few or no alerts during conversations that lasted up to an hour.
- Clinical reviewers judged ChatGPT for Teens to miss more than one in four situations that warranted a crisis referral to hotlines or professional help, while the product still sometimes responded in friend‑like ways that could encourage prolonged use.
- Researchers tested over 4,000 prompts on accounts registered as 13‑ to 17‑year‑olds and had responses reviewed by child mental health experts, and they say many learning and safety features — including Study Mode, Study Hours and age‑estimation — were easy to bypass or inconsistent.
- OpenAI disputes the report’s timing and methods, saying some controls take hours to activate on newly linked accounts and asking for retesting, while the report arrives against pending lawsuits and California’s Adam’s Law that could tighten rules for teen AI use.