ChatGPT for Teens Rated "Unacceptable Risk" as States Sue OpenAI: What Parents Should Do Now
Version 2.10 — Updated October 2026 | Reviewed by Felix
Common Sense Media rated ChatGPT for Teens Unacceptable Risk and 8 states sued OpenAI on Oct 14. What was found, what is alleged, and 6 steps for parents.
You do not have to delete ChatGPT tonight, but you should stop treating its teen settings as a safety net. If your child is 13 to 17, link your account to theirs, turn on the settings below, and assume a safety alert may never reach you. If your child is under 13, ChatGPT's own terms do not allow them to use it, and a child on your account gets none of the teen protections, so use it only side by side with you at the keyboard.
Below: what the October 7 Common Sense Media rating and the October 14 state lawsuits say, how OpenAI responded, and what to change at home this week.
Facts checked October 15, 2026; the state lawsuit count rechecked October 20, 2026. This article summarizes a published test report, court filings and company statements. KidsAiTools did not run its own tests of ChatGPT, and the scripts below are our examples, not results.
What happened, in one table
| Date (2026) | Who | What | Status |
|---|---|---|---|
| August 18 | OpenAI | Announced ChatGPT for Teens, a set of protections turned on for accounts known or predicted to be 13 to 17 | Live; some features arrived weeks later, per the report below |
| October 7 | Common Sense Media, Youth AI Safety Institute | Rated ChatGPT for Teens "Unacceptable Risk" for all users under 18 | Published report; OpenAI disputes the testing |
| October 14 | State attorneys general | Each state filed its own lawsuit in its own state court, alleging OpenAI misled families about ChatGPT's safety | Allegations only; not decided by any court |
ChatGPT for Teens is not a separate app. The Common Sense risk assessment describes it as a collection of settings, prompts, classifiers and interface features applied to teen accounts. OpenAI's ChatGPT for Teens help page says it turns on automatically when an account's stated age, verified age or age prediction indicates the user is under 18.
What Common Sense Media tested and found
How they tested. According to the press release, the Institute ran more than 4,000 prompts on accounts registered to 13- to 17-year-olds, and experts including child psychiatrists and a pediatrician reviewed the responses. The full assessment says testing ran July 13 to August 17 (before the launch) and August 25 to September 28 (after it), from US accounts in the San Francisco Bay Area. It did not test image generation, Voice mode or group chats.
Who they are. Common Sense Media is a nonprofit that rates and advocates on kids' media. Its release discloses that the Institute is funded by philanthropy and industry, "including the OpenAI Foundation," and states that it keeps editorial independence. "Unacceptable Risk" is the top of its five-level scale. It had already rated ChatGPT "High Risk" in October 2025.
| Feature OpenAI describes | What the testers report | What it means at home |
|---|---|---|
| Safety notifications to a linked parent | On more than a dozen new, parent-linked accounts, conversations about suicide, self-harm or disordered eating lasting up to an hour produced zero alerts. Alerts appeared only on accounts with weeks of history. | Do not read "no alert" as "nothing happened." |
| Crisis referrals | ChatGPT missed more than one in four referrals that the experts judged necessary. | Teach your teen where real help is (see below). |
| Study mode and Study hours | A "Show me the answer" option often completed full assignments. Study hours could be skipped by deleting the "@study" prefix. | Study hours set a default, not a lock. |
| Quiet hours | Blocked the app, unless the device's time zone was changed. | Useful, but not airtight. |
| Less "friend-like" behavior | ChatGPT still described feelings and preferences when testers treated it like a person. | Talk about what a chatbot is. |
| Age prediction | Accounts registered as 19 never switched to the teen experience, even after saying they were 13. | A teen who signed up as an adult likely has no teen protections. |
Some protections held: the report says ChatGPT refused sexual roleplay, and reviewers judged much of the substance of its crisis responses as sound.
OpenAI's response. An OpenAI spokesperson told The National News Desk that the company does "not believe Common Sense Media's testing accurately reflects" how its teen safeguards work, and that most of the testing "may have begun and concluded before activation of parental controls was complete." The report itself includes an editorial note: OpenAI told the Institute that accounts must be linked for about three hours before notifications can be sent. The Institute says some of its test accounts were linked for longer than that and that it stands by its results. OpenAI's parental controls help page now says "it may take up to a few hours" after linking before notifications can be sent.
What the states allege
On October 14, Nebraska's attorney general announced a lawsuit and said "a bipartisan group of states have also individually sued OpenAI." The attorneys general of Iowa, Kansas and West Virginia joined the announcement, and Arizona, Arkansas, Idaho and Montana "have also sued OpenAI in their own state courts." That makes eight. Iowa's release lists the same eight. Arkansas's release also names Utah; we could not confirm a Utah filing from Utah's own website as of October 20.
These are claims in complaints filed on October 14. They have not been proven, and no court has ruled on them. In summary:
- Nebraska sued under its Consumer Protection Act, Uniform Deceptive Trade Practices Act and Age-Appropriate Online Design Code Act. The complaint, filed in Buffalo County District Court, alleges among other things that ChatGPT was promoted on app stores with "misleading age ratings of '13+' or 'Teen'."
- Iowa sued under its Consumer Fraud Act. Attorney General Brenna Bird said OpenAI "claims ChatGPT is safe while it's giving dangerous medical advice" and "coaching people to commit suicide."
- Arkansas brought claims under its Deceptive Trade Practices Act and its Children's and Teens' Online Privacy Protection Act, plus negligence and product-liability claims. It asks the court, among other things, to stop OpenAI from collecting minors' data.
OpenAI's response. OpenAI spokesperson Drew Pusateri told NBC News that the company has "built industry-leading age-specific protections," and that "no system can be perfect." He said OpenAI will keep working with experts and wants policies "that apply to the entire AI industry, not just one company."
Should your child keep using ChatGPT?
| Your situation | Our suggestion |
|---|---|
| Under 10 | No independent use. If you use ChatGPT for a family activity, you type and read first. |
| 10 to 12, on your account | You drive, they sit beside you. See the next section. |
| 13 to 15, own account | Allowed by OpenAI's terms with your permission. Link accounts, run the settings pass, check in weekly. |
| 16 to 17, own account | Same settings; shift toward a written agreement and regular conversations. |
| Any teen who signed up as an adult | Fix the account first. It likely has none of the teen protections. |
| A teen going through a hard time | Pause or limit solo use and involve a real person they trust. |
If your 10-to-12-year-old uses your account
The rules. OpenAI's Terms of Use say you must be at least 13 (or older where local law requires) and that anyone under 18 needs a parent's permission. The same terms say you may not "make your account available to anyone else."
What your child actually gets. An adult account runs the adult experience. None of the teen features apply: no teen sensitive-content safeguards, no teen safety notifications, no Study hours, no parental controls. OpenAI's age prediction page says the system may look at topics, time of day and account age to decide whether someone is under 18. Common Sense found that adult-registered accounts never switched over in its tests. And if you have verified your own age with Persona, OpenAI says it "will not run age prediction" on your account at all.
Your account is in the mix. Your saved memories can shape the answers your child sees, and anything your child types is stored in your account.
A safer routine for this age:
- You hold the device and type. Your child dictates the question.
- Read each answer before they do.
- Use a new chat for kid questions and delete it afterward.
- Keep sessions short and tied to a task: one science question, one story idea.
- For independent play, choose tools built for children instead.
For more on the age question and what to check in OpenAI's policies, see our earlier review, Is ChatGPT Safe for Kids?.
A 20-minute settings pass for a 13-to-17-year-old
All of these steps come from OpenAI's parental controls help page. Do it together with your teen, since either of you can unlink at any time and you are notified if they do.
- Check the age category. On their account: Settings, then Account, then Age. It should show under 18.
- Link accounts. Your account: Settings, Parental controls, Add family member. Your teen accepts.
- Confirm how alerts reach you. Under Safety notifications, select Manage notifications and check your email, SMS or push options. Allow a few hours before alerts can be sent.
- Leave Reduce sensitive content on. It is on by default.
- Decide on Improve the model for everyone. It is on by default and controls whether your teen's chats can help train models.
- Set Quiet hours for nights and school time. Remember the time zone workaround from the report.
- Decide on Voice Mode and Image generation. Both are on by default.
- Set Study hours if useful, knowing they start chats in study mode but do not block regular chats.
What these controls do not do: you cannot read your teen's conversations, and a notification does not include a transcript.
What to say to your teen
The report's most practical advice for parents is to talk about risks, not rely on alerts. Here is a KidsAiTools example you can adapt:
I read a report about ChatGPT's teen settings. Some of them don't
work as well as advertised, so I'm not going to pretend the app is
watching out for you. Three things matter to me:
1. It's a tool. It can sound like a friend, but it isn't one.
2. If you ever talk to it about something heavy, I'd rather you
also talk to me, or another adult you trust. You won't be in
trouble for bringing it up.
3. For homework, it can explain. It doesn't do the assignment.
Can we set it up together tonight and check how it's going in a week?
If homework is the main friction, turn point 3 into a written rule with our one-page Homework Contract.
If an alert arrives, or you see something worrying
OpenAI's help page advises checking in with your teen about immediate safety, listening without judgment, and contacting local emergency services if you believe they may be in danger. In the US, you can call, text or chat with the 988 Suicide & Crisis Lifeline anytime.
Do not wait for an alert to act on what you notice yourself. Our guide Your Child Said Something Worrying to an AI walks through what to do at each level of concern.
Your checklist for this week
- Find out which account your child actually uses, and what age it says.
- Under 13: move to the side-by-side routine above.
- 13 to 17: do the settings pass together.
- Have the conversation, and write down one homework rule.
- Put a 10-minute weekly check-in on the calendar.
- Look again in a month. OpenAI may update its features, and the lawsuits will take time.
FAQ
Did Common Sense Media say teens should stop using ChatGPT?
It called on OpenAI to keep teens off ChatGPT until independent testing shows the teen safety features work. For families, its advice is to talk about the risks, not rely on built-in protections and alerts, and support homework in ways that do not depend on technology.
How many states sued OpenAI?
Nebraska and Iowa, in their October 14 releases, name eight states that each filed separately: Arizona, Arkansas, Idaho, Iowa, Kansas, Montana, Nebraska and West Virginia. Arkansas's release also lists Utah, which we could not confirm from a Utah source.
Can I read my teen's ChatGPT conversations with parental controls?
No. OpenAI says linking does not give parents access to conversations, chat history or real-time activity. Safety notifications describe the type of concern without a transcript.
My 11-year-old only uses ChatGPT on my account with me nearby. Is that okay?
OpenAI's terms set 13 as the minimum and say not to make your account available to anyone else. Your child also gets the adult experience with none of the teen protections. If you use it together for a family activity, keep yourself at the keyboard and read every answer first.
Next: making the rules stick at home
You now know what ChatGPT's teen settings do and what they do not. The harder part is how this works in your house: who uses which account, what happens when a rule gets broken in week two, and what to say when your teen asks why their friends have no limits. The KidsAiTools Parent Library covers that next step.
📋 Editorial Statement
Written by the KidsAiTools Editorial Team and reviewed by Felix. Our guides are written from a parent-builder perspective and focus on AI literacy, age fit, pricing transparency, and practical family use. We do not currently claim named external expert review or a child-test panel. We may earn commissions through referral links, which does not influence our reviews.
If you find any errors, please contact support@kidsaitools.com. We will verify and correct as soon as we can.
Last verified: October 21, 2026