Common Sense Media rated ChatGPT for Teens an “unacceptable risk” on October 7, saying its protections fall short of OpenAI’s promises. OpenAI disputes the assessment, arguing that parental controls may still have been activating during much of the testing. The nonprofit says it also saw missing alerts on accounts linked for longer.
The Verge reports that Common Sense Media’s assessment found failures in parental notifications and crisis responses. It also found that ChatGPT still completed children’s homework. OpenAI introduced the teen experience in August with guardrails and a focus on helping students learn.
Tom Siegel, executive director of Common Sense Media’s Youth AI Safety Institute, said the product could give parents misplaced confidence in its protections. He described a test in which a teen could discuss self-harm for an hour without a parent receiving an alert. Siegel called for ChatGPT to remain restricted to adults until OpenAI fixes the problems and demonstrates that through independent testing.
The dispute turns on activation time
OpenAI spokesperson Eric Porterfield told The Verge that the company welcomes independent evaluation but believes the testing misrepresents how its teen safeguards work. OpenAI said much of the testing may have started and ended before parental controls had finished activating. If so, the company argued, those tests would fail to establish whether notifications work as designed.
Common Sense Media stood by its methods. Siegel said the nonprofit had confirmed with OpenAI before testing that features including eating disorder notifications and Study Mode had launched.
For parental notifications, he said OpenAI disclosed after the tests that activation can take several hours on newly linked accounts. Some test accounts fell within that window. Others had been linked for substantially longer and still produced no notifications, according to Siegel. The nonprofit said it had shared its results with OpenAI before publication.
