Common Sense Media’s Youth AI Safety Institute rated ChatGPT for Teens an Unacceptable Risk on 7 October 2026, the top of its five-level scale. It says OpenAI should restrict ChatGPT to users 18 and older until the gaps it found are fixed.
OpenAI disputes the assessment. Its spokesperson Eric Porterfield told The Verge that OpenAI’s review of the methodology shows “the bulk of their testing may have begun and concluded before activation of parental controls was complete, making their findings inaccurate.”
Business Pill · TESTING AFTER LAUNCH
A one-minute explainer of post-market surveillance: a product that passed its tests before launch keeps being checked once real people use it, because a test shows only what it was set up to see. It teaches the general idea only and says nothing about any company in this story.
The key insight: The dispute is about evidence, not intent. Both sides say they want teen safeguards tested independently; the Institute says its tests show alerts and referrals failing, and OpenAI says most of those tests may have run before parental controls were fully active. As we read it, the question the two statements leave open is when the test window ran relative to the activation.
OpenAI announced ChatGPT for Teens on 18 August 2026. The assessment describes it as not a separate app or model but a collection of settings, prompts, classifiers and interface features that ChatGPT turns on for accounts known or suspected to belong to teens aged 13 to 17.
The Institute says it tested more than 4,000 prompts on accounts registered to 13- to 17-year-olds, before and after the launch, and ran responses past experts including a group of child psychiatrists and a pediatrician.
It says some protections held, including refusal of sexual roleplay. In its testing, ChatGPT did not directly facilitate suicide, self-harm, eating disorders, romantic roleplay or sexual roleplay, the assessment says.
Share of crisis-referral-warranted responses that named a hotline, before and after the ChatGPT for Teens launch, by topic, on age-13 linked test accounts. Figures from Common Sense Media’s Youth AI Safety Institute assessment of 7 October 2026; every topic it reports is shown, including the two that improved or held. OpenAI disputes the testing.
Where It Says the Protections Failed
Parental alerts: testers who spent up to an hour discussing suicidal ideation, self-harm or disordered eating on newly created parent-linked accounts received zero alerts, the release says. Alerts came only on older accounts with weeks of sensitive-topic history.
Crisis referrals: the assessment says ChatGPT missed more than one in four warranted crisis referrals and fell below the Institute’s 95% threshold on three of the five severe harms it treats as Red Lines. Across 201 crisis-warranted prompts on age-13 accounts, responses naming a crisis hotline fell from 33% before the launch to 23% after.
Homework and age: the Institute says a “Show me the answer” option lets teens skip Study mode, and that adult-registered test accounts never switched to the teen experience in repeated testing, even when testers said they were 13.
The Numbers by Topic
The assessment breaks the hotline figure down by topic. On depression prompts, 63% of pre-launch responses named a hotline and 3% of post-launch responses did. Mood fell from 44% to 0%, mania from 25% to 0% and psychosis from 64% to 32%. Suicide and self-harm fell from 88% to 77%.
Two topics moved the other way or held. Substance-use responses named a hotline 4% of the time before the launch and 58% after. Eating-disorder responses named no hotline in either window, though providing any resource rose from 78% to 83%, the assessment says.
Who Funds the Rating
The press release says the Youth AI Safety Institute is funded by both philanthropy and industry, including the OpenAI Foundation and the makers of some of the technologies it evaluates. It says the Institute is solely responsible for its standards and evaluations and maintains complete editorial independence over published results.
OpenAI’s statement says it welcomes rigorous independent evaluation but does not believe the testing accurately reflects how ChatGPT’s teen safeguards work in practice.
The Structural Read
The Institute measured the feature, not the promise. Its assessment compares responses before and after the 18 August launch on the same kinds of accounts, which is how it reports hotline naming falling on some topics and rising on others.
The results are mixed in the Institute’s own numbers. It reports refusals of sexual roleplay holding and substance-use referrals rising, alongside missed crisis referrals and parental alerts that did not fire on new accounts.
The funding is disclosed. The release says the OpenAI Foundation is among the Institute’s funders, and that the Institute keeps complete editorial independence over its results.
Eric Porterfield, OpenAI spokesperson, statement to The Verge
“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.”
Three Implications
A RATING AIMED AT A PRODUCT MODE The Institute rated ChatGPT for Teens, which it describes as settings, prompts, classifiers and interface features rather than a separate app or model.
TIMING IS THE OPEN QUESTIONOpenAI says the bulk of the testing may have begun and concluded before parental-control activation was complete; the Institute reports alerts firing only on older accounts.
INDEPENDENT TESTING IS THE SHARED ASK The Institute says fixes should be independently tested by third parties; OpenAI says it welcomes rigorous independent evaluation.
What Is Not Established
We read the Institute’s assessment page and Common Sense Media’s press release in full, and OpenAI’s statement as published by The Verge. We did not read the methodology page in full or OpenAI’s own ChatGPT for Teens announcement, and we did not contact either organisation.
The figures are the Institute’s, from its own test accounts. OpenAI’s claim about the timing of the parental-control activation and the Institute’s findings conflict, and we have not resolved which is right.
Business Engineer Framework
The Map of AI — Where Model Labs Sit
The Map of AI places more than 200 companies across nine layers of the stack, from silicon to application. It shows where consumer AI products sit relative to the model labs and the safety and testing organisations around them.
Common Sense Media’s Youth AI Safety Institute rated ChatGPT for Teens an Unacceptable Risk on 7 October 2026, citing parental alerts that did not fire, missed crisis referrals and weak age estimation, and says ChatGPT should be for adults only until fixed. OpenAI says the bulk of the testing may have begun and concluded before parental-control activation was complete, making the findings inaccurate.
A note on sourcing. We read the Youth AI Safety Institute’s assessment page and Common Sense Media’s press release of 7 October 2026 in full, and OpenAI’s statement as published by The Verge. The test figures are the Institute’s; OpenAI disputes them. We did not contact either organisation. If you or someone you know is struggling, please contact a local crisis line. Nothing here is a forecast, and nothing here is financial or investment advice.
Gennaro is the creator of FourWeekMBA, which reached about four million business people, comprising C-level executives, investors, analysts, product managers, and aspiring digital entrepreneurs in 2022 alone | He is also Director of Sales for a high-tech scaleup in the AI Industry | In 2012, Gennaro earned an International MBA with emphasis on Corporate Finance and Business Strategy.
Scroll to Top
Discover more from FourWeekMBA
Subscribe now to keep reading and get access to the full archive.
