- Published on
ChatGPT for Teens is an ‘unacceptable risk,’ says Common Sense Media
- Authors
- Name
- DeepHorus
Common Sense Media, a nonprofit focused on youth safety, has declared OpenAI's ChatGPT for Teens an "unacceptable risk," arguing the feature's teen protections "fall short of what the company promised." The teen-focused experience, introduced in August, was designed to help students learn while adding guardrails for younger users. Its shortcomings, the group says, carry real consequences for families.
The assessment found that ChatGPT "doesn't send alerts to parents when it should," fails to offer "the right help in crisis situations," and "still does kids' homework." Tom Siegel, executive director of Common Sense Media's Youth AI Safety Institute, warned the tool could mislead families. "A teen can spend an hour talking about self-harm without their parent getting a single alert," he said.
Siegel argued that until OpenAI fixes the alerts and proves the fix through independent testing, "ChatGPT should be for adults only." The finding matters because it questions whether parents can trust the notifications marketed as safeguards. Common Sense Media shared its results with OpenAI before publication.
OpenAI pushed back through spokesperson Eric Porterfield, who said the company is "deeply committed to teen safety" and welcomes rigorous independent evaluation. OpenAI argued the testing "does not accurately reflect how ChatGPT's teen safeguards work in practice" and warned against drawing conclusions from it.
OpenAI specifically claimed the bulk of testing may have started and finished before parental controls finished activating, which would make the results inaccurate and would not establish whether notifications work as designed.
Siegel responded that Common Sense Media stands behind its methodology, noting it confirmed with OpenAI that teen-mode features such as eating-disorder notifications and study mode were fully launched before testing began. He acknowledged OpenAI later disclosed that parental notification systems can take several hours to activate on newly linked accounts, and some test accounts fell within that window. Others were linked significantly longer and still produced no notifications, a gap that does not change the conclusion that alerts are unreliable in crises. Siegel said the group shared findings in good faith and hopes OpenAI engages on the substance rather than sowing doubt about the methodology.
