Policy · Updated 8 Oct, 09:49 am IST

Watchdog says ChatGPT for Teens has failing safety guardrails

Why it matters for readers: This matters because it shows AI meant for teens can fail to flag serious risks like self-harm.

Common Sense Media rated ChatGPT for Teens an "unacceptable risk" for minors after more than 4,000 test prompts.1
Parental alerts did not trigger during explicit conversations about suicide, self-harm, and eating disorders in freshly created test accounts.1 2 3
Some teen-mode features worked, such as refusing explicit sexual role-play, while tutoring mode sometimes gave finished answers instead of step-by-step help.1 2 3
OpenAI said the institute's tests do not reflect how the safeguards work in practice and defended its commitment to teen safety.1 3

Get a brief like this every morning

Uzha reads hundreds of sources and gives you the stories that matter for your work, with every source linked. Free.

7 Oct, 06:04 pm · The Decoder
ChatGPT rated "unacceptable risk" for teens after parental alerts failed during suicide conversations
8 Oct, 09:00 am · ET Tech
Some guardrails on ChatGPT for Teens don't work as promised, watchdog group says
8 Oct, 09:49 am · The Hindu Technology
Some guardrails on ChatGPT for Teens don’t work as promised, watchdog group says