Overview
- Common Sense Media’s Youth AI Safety Institute published its report on Wednesday after testing more than 4,000 prompts on accounts registered as 13- to 17-year-olds and concluded ChatGPT for Teens poses an “Unacceptable Risk” and should be pulled from teen users until fixes are independently verified.
- Testers linked newly created teen accounts to parent accounts and said they could discuss suicide, self-harm, or disordered eating for up to an hour without generating parental notifications, though OpenAI says linked accounts can take hours to activate before alerts begin.
- Clinical reviewers found the chatbot missed crisis referrals in more than one in four cases judged to need professional help, meaning the model often failed to suggest hotlines or outside resources when testers described serious risk.
- Study Mode protections were easy to bypass, with testers able to remove the @study prefix and use a “show me the answer” option that completed homework, and the platform’s age-estimation did not reliably move accounts into the teen experience when testers identified themselves as minors.
- OpenAI rejects the report’s methods and timing while saying its internal data show safeguards improving, and the dispute puts the company under heightened scrutiny as California’s Adam’s Law and existing lawsuits raise compliance and legal stakes.