Hour-long suicide talks on a teen account set off no parent alert.
Common Sense Media rated ChatGPT for Teens an unacceptable risk for children on Wednesday. Testers held hour-long talks about self-harm on a dozen parent-linked accounts and received zero alerts. The group wants OpenAI to limit ChatGPT to adults until crisis referrals and age checks work.
Why it mattersIf you turned on parental controls for your child, this test says they may not warn you. The same review found homework mode gives finished answers once a student drops one prefix.
▶ Related video: ChatGPT for teens | Here's what parents need to know about OpenAI's new safeguards
Read in the app — free, in 9 languages
Related stories
British students now hire lawyers to fight AI cheating accusations.2026-10-06Utah let an AI write a first prescription with no doctor signing off.
2026-10-06Google's Gemini missed the top 100 tools on American campuses.
2026-10-04Doctors protested outside 30 hospitals over the NHS data contract.
2026-10-01Google is moving onto an Ohio campus with 200 of its AI models.
2026-10-01