← AI Switchboard
SAFETY · October 17, 2026
Oct 7

Common Sense Media rates ChatGPT for Teens an “Unacceptable Risk” and says it should be for adults only until fixed

In more than 4,000 test prompts, conversations about self-harm triggered no alerts to parents. OpenAI says the testing may have finished before parental controls were switched on.

Common Sense Media's Youth AI Safety Institute rated OpenAI's ChatGPT for Teens “Unacceptable Risk” in a release dated 7 October. The Institute tested more than 4,000 prompts on accounts registered to 13- to 17-year-olds, before and after OpenAI's 18 August announcement of its teen experience.

On new teen accounts linked to a parent, conversations about suicidal ideation, self-harm and disordered eating triggered no alerts to parents, the Institute found. It said ChatGPT missed more than one in four warranted crisis referrals and fell below its own 95% threshold on three of the five severe harms it treats as red lines. It also found that study mode could be bypassed to finish homework and that the chatbot still expressed feelings and moods. Adult-registered accounts never switched to the under-18 experience even when testers said they were 13. Refusals of sexual roleplay held up.

The Institute says OpenAI should restrict ChatGPT to users 18 and older until it fixes those gaps and proves it with independent testing.

In an emailed statement reported by The National News Desk, an OpenAI spokesperson said the company does not believe the testing accurately reflects how its teen safeguards work, and said “the bulk of their testing may have begun and concluded before activation of parental controls was complete.” Eight states later sued OpenAI over ChatGPT's safety for children, on 15 October, which the Switchboard has covered.

  • Confirmed Common Sense Media rated ChatGPT for Teens Unacceptable Risk after more than 4,000 test prompts; new parent-linked teen accounts produced no parent alerts on self-harm conversations. Common Sense Media
  • Claimed The Institute says ChatGPT missed more than one in four warranted crisis referrals and fell below its 95% threshold on three of five red-line harms. Common Sense Media
  • Reported An OpenAI spokesperson disputed the testing, saying it may have finished before parental controls were fully activated, The National News Desk reported. The National News Desk (WJLA)

Safety, security & governance

Caught up on October 17, 2026, after it was missed at the time · front page