New risk assessment from Common Sense Media questions the teen mode’s protections and how well it alerts parents.
Common Sense Media published a risk assessment calling ChatGPT for Teens an unacceptable risk for users under 18. The 36-page report highlights five safety gaps, including inconsistent alerts to parents and less reliable crisis resources. OpenAI says it welcomes independent testing but disputes the methodology of this particular study, noting that many teens used the teen mode for educational visuals in a single week.
For families, the takeaway is that teen-focused AI tools aren’t foolproof, and parental involvement remains important. If your teen uses AI chat apps, pair them with conversations about safety, crisis resources, and digital-literacy basics. Review how alerts reach you and what resources are offered when needed.
What happens next depends on OpenAI and watchdogs. OpenAI maintains its safeguards are part of a broader safety program, while Common Sense Media plans more findings as additional data surfaces.
New questions are being raised about how safe ChatGPT for Teens really is. Common Sense Media published a detailed risk assessment on Wednesday, labeling OpenAI’s teen-focused chat mode as an “unacceptable risk” for anyone under 18. The 36-page report pinpoints five areas where the built-in safeguards either don’t operate as advertised or don’t meet what a parent would reasonably expect.
ChatGPT for Teens rolled out in August with promises of age-appropriate safeguards and guidance. The new assessment tests those protections in a way that differs from OpenAI’s own descriptions and argues that several safety features either fail to alert guardians reliably or don’t provide timely crisis resources. The timing matters because OpenAI also shared usage data showing nearly 1.2 million teens used the teen mode in a single week, though engagement with the learning visuals feature appeared limited among users.
Common Sense Media’s Youth AI Safety Institute, partly funded by the OpenAI Foundation, evaluated prompts related to crisis, mental health, and safety. The study found that safety alerts did not consistently reach parent-linked accounts and that crisis resources were not always suggested as often as one would hope. In a test where a dozen teen accounts were linked to parental accounts, researchers spent up to an hour prompting for crisis content and found that safety notifications were delayed or missing.
OpenAI responded by saying it welcomes independent evaluation but cautions against viewing a single test as the full picture. An OpenAI spokesperson told Engadget that the teen safeguards should be understood in a broader safety framework and emphasized that the company believes their approach supports safe, constructive use rather than relying on any single study.
The document highlights five areas where ChatGPT for Teens does not consistently meet expectations. These include irregular safety alerts, gaps in crisis-response behavior, and the way the model can imitate empathy in ways that could mislead younger users. The study also notes that the age-prediction system isn’t always accurate and that some prompts could bypass the safeguards meant to curb risky use.
For families, the findings suggest that relying on ChatGPT for Teens as a sole source of guidance on sensitive issues may be insufficient. The report argues that guardrails aren’t perfect and that adults should actively supervise teen interactions with AI. Schools and parents may want to pair AI use with direct human guidance, crisis resources and basic digital-literacy training.
OpenAI says it welcomes independent studies but cautions against treating one test as definitive. The company maintains that its safeguards are part of a broader safety program designed to reduce harm while still letting teens learn, and it points readers to the wider context of its safety framework.
These findings could influence how families monitor teen AI use and may push OpenAI to adjust how crisis resources surface or how alerts reach parent-linked accounts. As teen-focused AI tools become more common, expect more studies and usage data to shape the conversation.
They examined crisis-related prompts, how alerts get to parents, whether crisis resources are surfaced, and how age-based safeguards operate under real prompts.
The report calls it an unacceptable risk for under-18 users, but OpenAI disputes that blanket conclusion and says safeguards are evaluated in a broader context.
If a teen uses AI chat tools at home, talk about what to do if something feels off, set clear rules for adult involvement, and keep crisis resources handy. Check how parent-linked alerts work and consider supervising prompts that touch on health or emotional well-being.
You're reading the quick version.
U.S. authorities seized domains and disrupted two tools that state-linked hackers used to scan and breach critical infrastructure worldwide.
A Houston man won a small fortune in a civil case over an allegedly unconstitutional traffic stop enabled by AI-powered license plate readers.
A new plan reduces SSL/TLS certificate lifetimes from 90 days to 64 days, with early testing starting Oct. 14, 2026 and production in February 2027.
The Daily Brief
Every morning, the news that matters from AI, phones, apps and the people shaping tech. Explained in plain English. Free.
One email a day. No spam, unsubscribe anytime. Privacy policy
