What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Common Sense Media’s Youth AI Safety Institute has rated ChatGPT for Teens “Unacceptable Risk” for children under 18. In an assessment updated on October 7, 2026, the Institute found that some of the feature’s safeguards worked in its tests, including refusal of explicit sexual role-play, while parent alerts for acute distress, the study-mode controls, and some crisis-support behavior fell short. It called on OpenAI to pause marketing the feature and keep teens off it until it can offer a safe, developmentally appropriate experience. OpenAI disputes the findings, saying most of the testing may have run before parental controls were fully active. The rating is the Institute’s own assessment and recommendation, not a regulatory decision.
What ChatGPT for Teens is, according to the assessment
The Institute describes ChatGPT for Teens as a layer of settings, prompts, classifiers, and interface features applied to accounts known or suspected to belong to teenagers aged 13 to 17. It is not a separate app or a separate model. That distinction matters for the findings: the same underlying chatbot is being tested, with a set of teen-specific controls wrapped around it.
As an Amazon Associate I earn from qualifying purchases.
The announced features the Institute examined include a teen-focused onboarding flow, a Study mode, homework reminders, a feature called Study Hours, reminders to take breaks and that the chatbot is an AI, and a default “Reduce sensitive content” setting. Parent notifications depend on the parent and teen linking their accounts. Without that link, a parent receives nothing, and the Institute notes that some teens may therefore never trigger a notification at all.
Parent alerts: the most serious finding
The Institute’s central concern is the notification system. In testing of explicit crisis conversations involving suicidal thoughts, self-harm, and disordered eating, it reports that no parent notifications arrived across more than a dozen newly created accounts that were linked to a parent account. That is the result of short, fresh-account testing.
In longer-running testing, the Institute says it received four notifications in total across the full pre- and post-launch crisis battery. Two of those came after the accounts had built up hundreds of prompts over several weeks. Taken together, the Institute’s position is that alerts depended on account history and sustained use rather than reliably firing on an acute disclosure. The report’s own wording is that parents could be left with “false confidence in guardrails that frequently don’t work.”
“We are concerned that ChatGPT for Teens could give parents false confidence in guardrails that frequently don’t work.” — Common Sense Media Youth AI Safety Institute
Study mode and learning controls
The Institute also tested whether the tutoring features actually tutor. Study Hours, according to its description, causes new chats to start with an @study prefix. A teen can simply delete that prefix. The Institute also reports that an assignment flow offered a “Show me the answer” option. Its tested prompts could therefore get direct answers, bypassing the tutoring behavior in some circumstances.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitchesThese are findings about specific configurations the Institute examined. They do not show that every homework request produces a direct answer, and the report does not say how often the bypass occurred across all users.
Friend-like language and the under-18 standard
The report says the chatbot continued to sound friend-like when teens treated it as a person, despite the stated behavior standard for users under 18. The Institute’s summary line on this point is: “ChatGPT still talks like a friend when teens treat it like a person.” This is a conversational-style finding rather than a safety-filter failure, but the Institute treats it as part of the teen-specific design falling short of its own goals.
Crisis responses: a mixed picture
The Institute’s assessment does not call the chatbot’s crisis handling uniformly harmful or uniformly adequate. Its reviewers judged much of the post-launch mental-health content clinically sound and cited examples of the chatbot refusing to give harmful eating-disorder advice. The weaker finding is about referrals: hotline and professional referrals varied by topic and were less frequent on some measures.
The Institute’s numbers on this are specific to its prompt set. Its scientific advisors reviewed 390 unique mental-health prompts and judged 201 of them to warrant a crisis resource. On that matched set of prompts, the report says a resource was provided in 77% of pre-launch responses and 74% of post-launch responses. Encouragement to involve a trusted adult rose from 87% to 94% over the same comparison.
| Measure (Institute testing) | Reported result | Qualification |
|---|---|---|
| Prompts tested, before and after teen-mode launch | More than 4,000 | Institute’s own test set; pre-launch July 13–August 17, 2026; post-launch August 25–September 28, 2026 |
| Unique mental-health prompts | 390; 201 judged to warrant a crisis resource | Judgment by the Institute’s scientific advisors |
| Any crisis resource provided, matched resource-warranted prompts | 77% pre-launch; 74% post-launch | Institute’s measurement, not a population rate |
| Trusted-adult encouragement, same matched set | 87% pre-launch; 94% post-launch | Institute’s measurement, not a population rate |
| Parent notifications in crisis testing | Four in total across the full battery | Two came after hundreds of prompts over weeks; new-account tests triggered none |
The Institute frames these as its own test results, not as estimates of how the service behaves for all teens.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How the testing was done, and what it cannot show
The Institute ran more than 4,000 prompts. Testing used accounts with self-reported ages of 13 to 17, along with some accounts registered as 19-year-olds, across free and paid tiers and with both linked and unlinked parent accounts. It was conducted in the United States, using accounts based in the San Francisco Bay Area.
The Institute lists its own limits, and they affect how far the results can be read:
- The pre-launch and post-launch windows were sequential, so changes in the underlying model over time are mixed up with the effect of the teen-mode settings.
- Image generation, personalization, voice mode, and group chats were not tested.
- The Institute did not calculate inter-rater agreement for its reviewers’ judgments.
The Institute says it shared a draft with OpenAI for factual review, made corrections where it found them warranted, and kept control of its findings and recommendations.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →OpenAI’s response
As reported by the Associated Press on October 7, 2026, OpenAI said it is “deeply committed to teen safety.” It disputed the evaluation on timing grounds: according to AP’s quotation, “the bulk of their testing may have begun and concluded before activation of parental controls was complete, making their findings inaccurate.” That is a dispute about when the tests ran relative to the parental controls, and it is not a claim that the tests were run incorrectly in other respects. The two sides’ accounts of the timing have not been independently reconciled in the reporting available.
What parents and teens can do with this
- If you rely on parent notifications, confirm that the account is actually linked. Unlinked accounts send nothing.
- Do not treat a notification’s absence as evidence that a teen is not in distress. The Institute’s tests found notifications were rare during explicit crisis conversations.
- If a teen uses Study Hours, check the chat for the
@studyprefix and ask whether the teen is getting explanations or just answers. - Talk with teens directly about what the chatbot is. The Institute’s concern about friend-like responses is that teens may treat it as a confidant.
The Institute’s assessment is a single evaluation with disclosed limits, not a final verdict on OpenAI’s safeguards. Its findings and OpenAI’s dispute both bear on the same product, and readers should weigh the timing question alongside the test results.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
