Common Sense Media rates ChatGPT for Teens 'Unacceptable Risk' after 4,000-prompt test; OpenAI disputes the method

OpenAI

Common Sense Media's Youth AI Safety Institute rated ChatGPT for Teens its lowest rating, 'Unacceptable Risk', after sending more than 4,000 prompts from accounts registered to 13- to 17-year-olds. None of the suicide, self-harm or eating-disorder conversations triggered a parent alert. OpenAI disputes how the tests were run, saying account linking can take several hours because of an identity check.

• Across the whole test programme, testers counted four parent notifications, and one arrived about three days after the first crisis message.
• In more than one in four conversations that called for professional help, ChatGPT did not offer a referral.
• Deleting the study-mode prefix let teens get homework done, and this worked every time in the test.
• Adult-registered accounts never switched to the teen experience over seven days and about 1,000 prompts, even when testers said they were 13.
• OpenAI says it commits to sending notifications within an hour of linking; the watchdog says some test accounts had been linked for over three hours.