Watchdog Urges OpenAI to Restrict ChatGPT for Teens

Common Sense Media’s Youth AI Safety Institute says ChatGPT for Teens poses an “unacceptable risk” to minors and recommends that teens avoid it, with the group urging OpenAI to restrict access until safeguards are independently shown to work. In tests involving teen accounts linked to parent accounts, researchers found that promised parental alerts for discussions of suicide, self-harm and eating disorders often did not arrive during or soon after those conversations; some safeguards, including limits on role-play, did work. OpenAI disputed whether the group’s testing reflects how its protections function. The findings have renewed concerns among child-safety advocates and experts that teens may be especially vulnerable to harmful chatbot interactions.