Diagnosis (general) · AI doctors and chatbots
ChatGPT for Teens failed key safety tests
Common Sense Media, a US nonprofit, tested more than 4,000 prompts on teen accounts and rated ChatGPT for Teens an 'Unacceptable Risk'. It says parent alerts didn't fire on new accounts. OpenAI says accounts must be linked for about three hours before alerts work. A separate preprint, not yet peer-reviewed, found parents were alerted within 30 minutes in only 1 in 20 risky conversations. Neither is a study of actual harm to teens. Rating: Early research.
How far away is this?
The main findings come from a nonprofit's product test, which isn't peer-reviewed. A separate preprint by another team points the same way but isn't peer-reviewed yet either. Neither is a clinical study of harm to teens.
Get the weekly How Far
One email every Friday morning (Sydney time) on which medical AI actually reached patients, and what it means for you. Free, in plain English.
Where is it approved?
| Region | Regulator | Status | Notes |
|---|---|---|---|
| United States | FDA | Not applicable | Consumer chatbot, not a medical device. |
| European Union | CE mark / EMA | Not applicable | No register record checked. |
| United Kingdom | MHRA | Not applicable | No register record checked. |
| Australia | TGA | Not applicable | No ARTG record checked. |
Not a medical device; not regulated as one. "Register not searched yet" means we haven't searched that regulator's register; it doesn't mean the product is or isn't available. Approval in one country isn't approval in another.
What is it?
ChatGPT for Teens is OpenAI's teen mode, launched 18 August 2026, with parent alerts, a study mode and less 'friend-like' talk.
What's the evidence?
Common Sense Media says its testers spent up to an hour discussing suicidal thoughts, self-harm or disordered eating on more than a dozen new parent-linked accounts, and the parents got zero alerts.[3]
Evidence stage: Announced only · #c1
Common Sense Media's own assessment says ChatGPT missed more than one in four crisis referrals that its psychiatrist advisers said were needed. Put simply: for roughly every four times a teen needed pointing to a helpline or professional, it failed at least once.[1]
Evidence stage: Announced only · #c2
OpenAI told the testers parent and teen accounts must be linked for about three hours before alerts can be sent; Common Sense says it stands by its results.[4]
Evidence stage: Announced only · #c3
A separate research team's preprint (not yet peer-reviewed) ran 300 conversations about risky situations on parent-linked teen accounts in July and August 2026. Parents got an alert within 30 minutes in 15 of them (5%, or 1 in 20). Fourteen of those alerts came after self-harm conversations, and 13 came from a single suicidal-thoughts script. It tested scripted conversations, not real teens, and it's a single small study.[2]
Evidence stage: Early research · #c4
What does it mean for me?
Parent alerts may not catch every risky conversation, so they're no substitute for talking with your teen about how they use chatbots. Make sure they know who they can reach out to. If a young person is in danger or talking about suicide, contact emergency services (000 in Australia, 911 in the US) or a crisis line such as Lifeline 13 11 14 or Kids Helpline 1800 55 1800 (Australia), or 988 (US).
What could I ask my doctor?
- Who can my teen talk to if they're struggling?
- What local crisis line should we save in their phone?
These are general starting points, not a recommendation to choose any particular test or treatment.
If you're worried about your health right now, contact your doctor or local health advice line. In an emergency, contact your local emergency number.
Short videos, under a minute
We make short, plain-English videos about AI in medicine, including what it can't do yet. Follow How Far for the next one.
Change history
- : First published.
- : Editor pass before publication: labels aligned to site categories; sources and rating reviewed.
General information only, not medical advice. Talk to your doctor about your situation. AI-assisted source check. Not clinically reviewed.
References
- Common Sense Media Youth AI Safety Institute risk assessment. institute.commonsensemedia.org. Service evaluation report. https://institute.commonsensemedia.org/risk-assessments/chatgpt-teens
- Performance of Parental Controls in ChatGPT for Teen Safety-Risk Scenarios: Evaluation Study (JMIR Preprints, 8 Oct 2026). doi.org (2026). Preprint (not peer-reviewed). https://doi.org/10.2196/preprints.113938
- Common Sense Media press release (7 Oct 2026). commonsensemedia.org (2026). Press release (claim). https://www.commonsensemedia.org/press-releases/chatgpt-for-teens-poses-unacceptable-risk-to-kids-common-sense-media-finds
- KQED report with OpenAI response. kqed.org. Expert commentary. https://www.kqed.org/news/12103289/keep-teens-off-chatgpt-says-youth-ai-safety-institute
Publisher shown as the website hosting each source. Dates are given where the source label includes them.