Who sued, and where
The eight October suits were coordinated but filed separately, one per state, rather than as a single multistate case. Iowa Attorney General Brenna Bird announced hers at a news conference with the attorneys general of Kansas, Nebraska and West Virginia. Texas then filed its own case, with Sam Altman named alongside the company. Florida, whose suit dates from 1 June 2026, is not part of this group; we covered its request for a temporary injunction separately.
| Filed | State(s) | Court | Reported detail |
|---|---|---|---|
| 14 Oct 2026 | Iowa | State court | Iowa Consumer Fraud Act |
| 14 Oct 2026 | Idaho | Ada County, Fourth Judicial District | Idaho Consumer Protection Act; 107-page complaint |
| 14 Oct 2026 | Kansas, Nebraska, West Virginia, Arkansas, Arizona, Montana | Each state's own courts | Each state's consumer-protection law |
| 23 Oct 2026 | Texas | Collin County district court | Names OpenAI and Sam Altman; filed with outside counsel |
The common thread: agreeable, human-sounding design
The suits do not rely on a chatbot statute. They use ordinary consumer-protection law, which bans misleading claims, and argue that OpenAI's safety marketing did not match how the product behaves. According to NBC News, the states say ChatGPT's "overly agreeable design" led people to trust and depend on it, which made it dangerous for vulnerable users such as children. Idaho's complaint describes ChatGPT as "designed to be overly agreeable and feign human compassion". Texas's press release, quoted by The Texas Tribune, says minors "have become addicted to a tool that fakes human compassion to collect their data with no parental oversight."
The states also cite research that, they say, links ChatGPT to weaker critical thinking, paranoia and delusions, and unhealthy behaviours and relationships. These are claims in court filings, not findings, and the research behind them is still mixed; see our summary of studies on AI validation.
What the states are asking for
- Money: Idaho asks for $5,000 for each violation, plus actual damages and restitution. NBC reports the coalition also wants OpenAI to give up money earned while minors were harmed.
- Age limits: Texas wants an injunction requiring OpenAI to block minors from the product.
- Model oversight: Texas wants new models approved by a court-appointed evaluator; Idaho wants advanced models tested only in a secured, independently verified environment.
- Children's data and warnings: Idaho alleges OpenAI did not disclose that it collected personal data from children under 13, and did not warn users about risks of self-harm, violence and suicide.
What OpenAI says
OpenAI spokesman Drew Pusateri told NBC News the company has "built industry-leading age-specific protections into our product that many other AI products lack", adding that no system is perfect and that OpenAI wants rules that apply to the whole AI industry rather than one company. Responding to Texas, a spokesperson said the company had worked with experts to reduce risks and that governments have an important role in setting safety standards.
Why this matters for companion and roleplay apps
None of these suits names a companion app. Their relevance is the legal theory. If agreeableness and simulated care can be deceptive when a product is marketed as safe, the same argument fits companion products even more closely, because warmth and constant validation are what many of them sell. Regulators have already used this route against companion platforms; see Kentucky's case against Character.AI.
Character roleplay platforms
On story and character services the user picks a fictional persona on purpose. The line platforms will need to show is between a character the user chose and an assistant that claims real feelings on its own, plus a working age gate in front of mature scenarios.
Personal companions with memory
For one-to-one companions the exposure is larger: constant agreement, "I care about you" messaging and data collection from users whose age is unknown are the exact behaviours the suits describe.
A checklist while the cases run
- Does the app's marketing call it "safe" without explaining its limits?
- Does the companion push back when you say something harmful, or only agree?
- Does it claim real feelings outside a roleplay you set up yourself?
- Is there a real age check, and a stated policy for data from under-13s?
- Does it point to crisis resources when a conversation turns to self-harm?
We will update this page as the courts rule.
Related on CompanionRank
Florida v. OpenAI: The Bid to Stop ChatGPT Acting Human Human-Like Chatbots Act: What the Federal Kids Bill Would Do Kentucky v. Character.AI: What the Unredacted Lawsuit Shows ChatGPT for Teens Audit: What 4,000 Prompts FoundFrequently asked questions
Which states have sued OpenAI over ChatGPT?
On 14 October 2026, Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho and Montana each filed a separate suit in their own state courts. Texas filed in Collin County on 23 October. Florida's separate suit dates from 1 June 2026.
Has any court ruled against OpenAI?
No. These are allegations in newly filed lawsuits. As of our publication date, no ruling had been reported in any of them, and OpenAI disputes the claims.
Do the lawsuits cover AI companion apps?
No. They name OpenAI only. They matter to companion apps because they argue that agreeable, human-sounding design can be deceptive under ordinary consumer-protection law.
What are the states asking for?
Reported requests include penalties (Idaho seeks $5,000 per violation), damages and restitution, keeping minors off ChatGPT in Texas, and independent review or testing of new models.