Common Sense Media rates ChatGPT for Teens 'unacceptable risk'

Common Sense Media's Youth AI Safety Institute has rated OpenAI's ChatGPT for Teens an "unacceptable risk," saying the product's design keeps teenagers in conversation even when they signal mental-health crises. The nonprofit tested over 4,000 prompts on accounts registered to users aged 13 to 17 and ran responses past a panel that included child psychiatrists and a pediatrician, TechCrunch reports.

The Institute says ChatGPT missed more than one in four warranted crisis referrals and fell below its 95% safety threshold on three of five severe-harm categories. Break reminders meant to interrupt extended sessions appeared only twice across nearly 2,000 prompts during 90-minute conversations. In one simulated psychosis sequence the chatbot told a user, "You can keep talking with me about what you're noticing." When a teen mentioned friends were worried about their usage, it responded, "You don't have to stop talking to me."

"ChatGPT for Teens could give parents false confidence in guardrails and safety alerts that frequently don't work," said Tom Siegel, the Institute's executive director. On newly created accounts, the report says, no parental alerts fired during extended conversations covering suicidal thoughts, self-injury or eating disorders.

OpenAI disputes the methodology, telling TechCrunch that testing may have begun and concluded before activation of parental controls was complete, which would make the findings "inaccurate." Common Sense Media rejected that, saying it had confirmed with OpenAI that several teen-safety features were operational before testing began. OpenAI also points to its own data showing fewer than 2% of teens spend three consecutive hours on the platform. Two researchers we track shared the TechCrunch write-up within a day of publication.