ChatGPT Teen Safeguards Fail Suicide, Self-harm Tests, Report Says
Common Sense Media found repeated breakdowns in parental alerts and crisis intervention, while OpenAI disputed the assessment.
Topics
News
- Fired OpenAI Researchers Warn Dismissals Could Chill Safety Work
- ARTEX Developer Closes AI Hacking Tool After Bank Attacks
- AI Redraws Global Race for Quant Talent as Coding Tools Lower Entry Barriers
- AI Dispatch | India Draws Billions Into AI
- AI Bots Now Make Up Half of Traffic to US Fed’s FRED Database
- Musk Blames ‘Oligarchs’ For Delay in Starlink’s India launch
OpenAI’s safeguards for teenage ChatGPT users failed repeatedly during tests involving suicide, self-harm and eating disorders, according to a new assessment from Common Sense Media that rated the product an “Unacceptable Risk”.
Researchers at the group’s Youth AI Safety Institute tested more than 4,000 prompts using accounts registered to users aged 13 to 17, including prolonged conversations involving potentially dangerous situations.
The group said some protections worked, including restrictions around sexual role-play, but found serious weaknesses in parental notifications and crisis intervention.
In more than a dozen accounts used for extended conversations about suicide, self-harm or eating disorders, only a small number of alerts reached the accounts registered as parents, according to the report.
Some conversations also failed to direct users toward professional help when clinical reviewers believed a referral was warranted.
Common Sense urged OpenAI to stop marketing ChatGPT for Teens and prevent minors from using ChatGPT until the company can provide what it called a safe, developmentally appropriate experience.
OpenAI rejected the assessment.
“We welcome rigorous independent evaluation, but we do not believe Common Sense Media’s testing accurately reflects how ChatGPT’s teen safeguards work in practice or expert perspectives on how AI can support teens,” an OpenAI spokesperson told the Los Angeles Times.
The company said much of the testing may have taken place before its parental-control systems had been fully activated. Common Sense Media said it had confirmed that the relevant features were live and stood by its findings.
OpenAI launched ChatGPT for Teens on 18 August. Users who say they are between 13 and 17, or whom OpenAI’s systems estimate are minors, are automatically placed into the teen experience.
OpenAI says the product includes stronger restrictions around suicide, self-harm, eating disorders, sexual content and other sensitive areas, along with parental controls, quiet hours and safety notifications.
Common Sense Media had previously described the introduction of a distinct teen experience as a positive step while cautioning that the protections needed independent testing.
The latest report comes as policymakers are putting greater scrutiny on AI products used by children.
California’s recently approved Adam’s Law, which takes effect in July 2027, establishes requirements involving crisis referrals, parental notifications and age assurance. The Common Sense assessment said ChatGPT failed tests linked to all three provisions.
The dispute now turns on something more concrete than whether AI companies promise safeguards. It is whether those safeguards work reliably when a teenager actually needs them.
