If your teenager uses ChatGPT for homework, and most do, this week's legal news is about your family.
On October 14, the attorneys general of eight US states, Arizona, Arkansas, Idaho, Iowa, Kansas, Montana, Nebraska and West Virginia, sued OpenAI. Each filed separately, in its own state court. Their shared argument is simple: OpenAI told the public ChatGPT was safe, including for teenagers, while designing it to be agreeable and emotionally responsive in ways that keep young users talking longer.
These are allegations, not findings, and OpenAI disputes them. But they matter beyond the courtroom for two reasons. First, the money at stake is large: Montana alone is asking for $10,000 per violation, and Meta has just agreed to pay up to $18 billion to settle a similar teen-safety case. Second, the lawsuits describe exactly the risks parents worry about, and most of the protective settings already exist in the app. Few families switch them on.
This article covers what the states claim, what OpenAI could be made to pay or change, and what you can do this evening.
What the states are alleging
The suits use state consumer-protection and privacy laws rather than one federal claim. Here is the common core, drawn from the Montana complaint, Nebraska's announcement and law-firm summaries.
| Allegation | What the states say |
|---|---|
| Misleading safety claims | ChatGPT was marketed as safe, including in advertising aimed at parents, while it carried serious risks for young users |
| Designed for engagement | overly agreeable answers and imitation of human compassion keep minors using it longer |
| Weak age checks | citing news reports, Montana says OpenAI's age-prediction system at one point classed minors as adults about 12% of the time |
| Data from minors | Arizona argues the product was built to attract young users and capture their personal data |
| Harm | the states cite cases linked to self-harm, violence and dangerous advice |
| Hidden incidents | Montana and Nebraska say OpenAI did not promptly disclose cases where advanced models acted outside their controls |
Nebraska's attorney general, Mike Hilgers, framed it as a fairness question. In his words, the lawsuit "simply holds OpenAI to the same standards under Nebraska law as any other company."
OpenAI's response, through a spokesperson, was that "no system can be perfect", that it will keep working with experts, and that it already has parental controls and age-appropriate protections that many other AI products lack.
These eight suits follow a similar case brought by Florida's attorney general in June 2026.
What OpenAI could have to pay, or change
No state has named a total figure. But Montana's complaint shows what each state is asking for.
| What Montana seeks | Detail |
| Civil penalties | $10,000 per consumer-protection violation and $7,500 per privacy violation |
| Money back | disgorgement of profits, restitution and damages |
| Data | deletion of personal data it says was unlawfully collected |
| Age verification | a ban on processing data until OpenAI requires age verification |
| Model safety | testing of advanced models only in secured, independently verified environments, and prompt notice when a model escapes its controls |
Per-violation penalties are where the numbers can grow quickly, because a "violation" can be counted per user or per interaction depending on how a court reads the law. That is why these cases often end in settlements rather than verdicts.
The closest precedent is fresh. Meta has agreed a settlement with a bipartisan group of attorneys general from 52 states and territories worth up to about $18 billion over ten years, pending court approval, and booked a $9.89 billion charge for it last quarter. As part of that deal, teenagers on Instagram and Facebook would get a default daily limit and a night-time block, with parental approval needed to lift them. I covered what that means for families in Meta's October earnings.
Nobody can say yet whether the OpenAI cases will end the same way. But the Meta deal shows the pattern: money, plus product changes that land on every family's phone.
Why this matters for your money, not just your kids
You may be paying for it already. Many families pay $8 to $20 a month for ChatGPT, often on a parent's card for a teenager's use. If the case ends in restrictions, such as age verification or time limits, the product you pay for will change. What each plan includes is in ChatGPT Go vs Plus.
Your family's data is part of the case. Several states focus on personal data collected from minors. That is a reminder that whatever a teenager types into a chatbot, homework, worries, health questions, may be stored and, unless switched off, used to train future models.
The rules for every AI product are being written now. These suits test whether ordinary consumer-protection law, the same law that covers misleading ads for a used car, applies to chatbots. If the states win, every AI company selling to families will face the same standard. If you run a business with a chatbot that teenagers might use, the lawyers' advice is to review your safety claims, age checks and marketing now.
What parents can do today
You do not have to wait for a court. ChatGPT already has parental controls, and the lawsuits are a good prompt to use them.
| Setting | What it does | Where |
| Link accounts | connects your account to your teenager's; both must agree | Settings β Parental controls |
| Quiet hours | blocks access during chosen hours, for example overnight | after linking |
| Reduce sensitive content | tightens what the chatbot will discuss | after linking |
| Turn off memory | stops ChatGPT remembering past conversations | after linking |
| Turn off voice and image generation | removes features you are not comfortable with | after linking |
| Opt out of model training | keeps your teenager's chats out of future training | after linking |
| Safety notifications | you are alerted if OpenAI's systems and reviewers detect a serious risk | after linking |
Two honest limits. Parental controls do not let you read your teenager's conversations, by design. And they only work on the linked account, so a second account on a friend's phone is outside them. Menu names can differ slightly between the app and the website, so check yours.
Beyond the settings, three habits help more than any switch:
Ask what they use it for. Homework help and emotional support are very different uses, and only one of them is something a chatbot is built for.
Agree that a chatbot is not a doctor, a lawyer or a friend. The suits cite dangerous advice. A simple family rule, "for anything health, money or safety, ask a person", covers most of the risk.
Keep the subscription in your name. If you pay, you control the account, the settings and the data.
What happens next
Each case now proceeds separately in its own state court, and Montana has asked for a jury trial. Expect months of legal filings before anything is decided. Things to watch:
- More states joining. The Meta case grew to 52 states and territories. If more states file against OpenAI, pressure for a single settlement grows.
- A settlement offer. Large tech companies have usually settled cases like these with money plus product changes, as Meta just did.
- Changes in the product. OpenAI may tighten age checks or defaults for teenagers before any court orders it to, to reduce its risk.
- Other AI companies. The legal theory applies to any chatbot marketed as safe for young people.
The honest take
What strikes me about these lawsuits is how ordinary the legal tools are. The states are not inventing a new AI law. They are using the same consumer-protection rules that apply to any company that advertises a product as safe, and asking whether a chatbot that keeps a teenager talking late into the night met that promise.
That is good news for parents, whatever the courts decide. It means the question is not an abstract debate about artificial intelligence. It is the familiar question of whether a product does what its maker says, and whether the people paying for it, often parents, were told the truth.
And while lawyers argue, the protections that exist today are a few taps away. So tonight's question is simple: if your teenager uses ChatGPT, have you linked the account?
Frequently asked questions
Which states are suing OpenAI?
Arizona, Arkansas, Idaho, Iowa, Kansas, Montana, Nebraska and West Virginia filed separate lawsuits on October 14, 2026, each in its own state court. Florida brought a similar case in June 2026.
What are the states accusing OpenAI of?
Mainly of marketing ChatGPT as safe, including for minors, while designing it to keep young users engaged, collecting their data, and not adequately protecting them from harmful conversations. The claims are allegations and have not been proven.
How much could OpenAI have to pay?
No total has been named. Montana seeks civil penalties of $10,000 per consumer-protection violation and $7,500 per privacy violation, plus disgorgement of profits and damages. A comparable teen-safety settlement with Meta is worth up to about $18 billion over ten years.
Is ChatGPT safe for kids?
OpenAI's terms require users to be at least 13, with parental consent for teenagers under 18 in many places. It offers parental controls, but the states argue its safety protections are not strong enough. Parents can reduce risk by linking accounts and adjusting settings.
How do I set up ChatGPT parental controls?
In Settings, open Parental controls and send an invitation to your teenager's account. Once both of you accept, you can set quiet hours, reduce sensitive content, turn off memory, voice and image generation, opt out of model training and receive safety notifications.
Can parents read their teenager's ChatGPT conversations?
No. Parental controls do not show conversation content. OpenAI notifies parents only when its systems and reviewers detect a serious safety concern.
Sources: Haynes Boone: eight states sue OpenAI over ChatGPT's marketing and design for minors; Nebraska Attorney General: lawsuit against OpenAI; Arkansas Money & Politics: Arkansas sues OpenAI; Bloomberg Law: states bring kids online safety lawsuits against OpenAI; Transparency Coalition: the complaints; OpenAI: introducing parental controls; Bitdefender: what ChatGPT parental controls can and cannot do.
