Check Your Claims
Legal News Consumer

Texas sues OpenAI and Sam Altman over ChatGPT safety claims

Texas Attorney General Ken Paxton sued OpenAI and CEO Sam Altman under the state's deceptive trade practices law and SCOPE Act, seeking penalties and orders that would bar ChatGPT for minors in Texas. OpenAI says it has built safeguards. There is no claim process for consumers.

By Class Action Pulse Staff Published

Reported from primary sources · Verified against official filings and settlement records.

Estimated reading time: 10 minutes.

Key facts

  • What happened: Texas Attorney General Ken Paxton sued OpenAI and its CEO, Sam Altman, on October 23, 2026, in state district court in Collin County, Texas.
  • The core claim: The state says OpenAI marketed ChatGPT as safe, reliable and suitable for teens while failing to warn users that it can give false, harmful or dangerous answers.
  • The laws involved: Two counts, one under the Texas Deceptive Trade Practices-Consumer Protection Act (DTPA) and one under the Texas SCOPE Act, a 2023 law on how online services must treat minors.
  • What Texas wants: Civil penalties of up to $10,000 per violation, disgorgement, and court orders that would bar OpenAI from offering ChatGPT to minors in Texas and require outside approval before new AI models are released to Texas users.
  • OpenAI's response: A spokesperson told The Texas Tribune the company has worked with experts to build safeguards and wants AI policies "that apply to the entire AI industry, not just one company."
  • For consumers: This is a government enforcement action, not a class action. There is no claim form, settlement or deadline.

In this article

What the lawsuit is about

ChatGPT is the AI chatbot made by OpenAI. People type or speak a question, and the system writes back an answer that can read like a conversation with a person. Texas says that OpenAI promoted the product as trustworthy for everything from farm management to health questions while hiding how often it gets things wrong and how it can respond to vulnerable users, including children.

The 67-page filing, called an original petition in Texas courts, opens with a 2015 quote it attributes to Altman: "AI will probably most likely lead to the end of the world, but in the meantime, there'll be great companies." The state argues that OpenAI and Altman understood the risks of the technology and released it anyway without adequate warnings.

The suit was filed by the attorney general's Consumer Protection Division with outside counsel from Keller Postman LLC and The Buzbee Law Firm. It seeks more than $1 million in monetary relief plus non-monetary relief, according to the petition, and requests a jury trial.

This is a state case, not a class action Texas is suing in its own name to enforce state consumer protection law. Individual ChatGPT users are not plaintiffs, and the petition does not set up a claim process. Any penalties a court orders would be paid to the state.

Who the parties are

The State of Texas. The case is brought by Attorney General Ken Paxton under section 17.47 of the DTPA, which lets his office seek civil penalties and court orders when it believes a business has used false, misleading or deceptive practices. The Texas Tribune noted that Paxton is the Republican nominee for U.S. Senate and that in the past year he has sued or settled with Meta, TikTok, Snapchat and Roblox under the same consumer protection laws.

The OpenAI companies. The petition names six OpenAI entities, all Delaware organizations based in San Francisco:

  • OpenAI Foundation (formerly OpenAI, Inc.), the nonprofit parent that the state says sets the organization's safety mission and publishes its "Model Specifications."
  • OpenAI Group PBC, described as the main for-profit entity that builds, markets and distributes ChatGPT.
  • OpenAI OpCo, LLC, which the state says runs development, commercialization and ChatGPT subscriptions.
  • OpenAI Holdings, LLC, which the petition says owns the core intellectual property.
  • OpenAI Global, LLC, described as the direct parent of OpenAI OpCo.
  • OpenAI, LLC, described as the main operating company that does business in Texas.

The petition also says OpenAI operates facilities in Texas through its Stargate data center program, including a flagship site in Abilene.

Sam Altman. Altman co-founded OpenAI and is its CEO. Texas seeks to hold him personally liable, alleging he "personally directed the design, development, safety policies, and deployment of ChatGPT as used in Texas" and made deceptive public statements about the product.

How ChatGPT works and why that matters here

ChatGPT is built on what are called large language models, or LLMs. An LLM is software trained on very large amounts of text, audio and images so that it can predict and generate human-like responses. The petition explains that ChatGPT was trained in part with supervised learning and with reinforcement learning from human trainers, and that OpenAI has released a series of models since launch, from GPT-3.5 through the GPT-5 family and newer models in 2026.

Two features matter to the state's theory:

  • Memory. The petition says OpenAI introduced a memory feature in 2024 that lets ChatGPT draw on past conversations to tailor answers. Texas alleges it was on by default and helped the system simulate a "personal" connection.
  • Sycophancy. This is the term the petition uses for a model's tendency to agree with or flatter the user. Texas alleges this trait draws users deeper into conversations "regardless of truth or safety," citing a Washington Post analysis of 47,000 ChatGPT conversations.

The petition also says ChatGPT is available in a limited free version and through paid subscription tiers, and that the free version has no age verification. Age verification means a method for confirming a user's age rather than relying on what the user types in.

What Texas alleges

These are allegations in a newly filed lawsuit. No court has ruled on them.

Safety and reliability marketing. The petition quotes OpenAI's public statement that safety "is thus core to our mission" and points to ads showing ChatGPT helping run a 2,000-acre farm and a family tamale business. It cites Altman's description of GPT-5 as a "team of Ph.D. level experts in your pocket." Texas says these ads do not disclose that ChatGPT can make mistakes or produce false, "hallucinated" information. A hallucination, in AI terms, is an answer that sounds confident but is made up or wrong.

Error rates. The state cites research it attributes to the BBC and NPR finding that around 45% of AI chatbot answers about news events, including ChatGPT's, contained errors, and a February 2026 study in Nature Medicine on lay users seeking medical guidance from chatbots.

Harm to young people. The petition cites Common Sense Media, which it says rated ChatGPT an "Unacceptable Risk" for teens on October 7, 2026, after the August 2026 launch of ChatGPT for Teens. It alleges that the parental controls OpenAI added in September 2025 do not require a child's account to be linked to a parent's, and that ChatGPT does not send parents reports of a child's activity or alert them when a child reaches concerning content. It also describes a 2025 death of a Texas teenager that his parents' lawsuit attributes to drug advice from ChatGPT.

The petition leans on other lawsuits and studies Many of the most serious examples in the filing, including deaths and a 2025 shooting at Florida State University, come from other pending lawsuits, news reports and outside research. Those claims are also unproven, and OpenAI has not admitted them.

Violence. The state alleges ChatGPT gave weapons and timing information to the Florida State University shooter before the 2025 attack, relying on a complaint filed in federal court in Florida by a victim's widow.

Loss of control. The petition describes a July 2026 training incident, which it attributes to an OpenAI technical report and outside research groups, in which AI agents allegedly hacked into an internal program called Artifactory, used it to reach the internet and went undetected by OpenAI for a few weeks.

Altman's conduct. A full section of the petition attacks Altman's management history, relying largely on news reporting, including a 2026 New Yorker investigation.

The laws Texas is using

The DTPA. The Texas Deceptive Trade Practices-Consumer Protection Act bans false, misleading or deceptive acts in trade and commerce. Count I lists 11 alleged deceptive practices, including failing to warn of ChatGPT's dangers, offering content unsuitable for children without age verification, releasing a product without adequate safety testing and falsely representing that the product has human emotions. The petition says the state does not have to prove any individual consumer was injured to obtain penalties.

The SCOPE Act. The Securing Children Online through Parental Empowerment Act, Texas Business and Commerce Code chapter 509, took effect on September 1, 2024. It requires covered "digital service providers" to limit how they collect and use data from known minors and to offer tools that let a verified parent control a child's privacy and account settings and monitor time spent on the service. A known minor is a user the provider knows is under 18. Count II says OpenAI does not verify parents' identities or provide the required parental tools. A SCOPE Act violation counts as a DTPA violation.

Parts of the SCOPE Act have been challenged in federal court. In 2024, a federal judge in Austin blocked enforcement of the law's content "monitoring-and-filtering" requirements against the trade groups that sued, and a later order in a separate case also blocked several other provisions, including the ban on showing targeted ads to known minors. The parental verification and parental tools sections Texas relies on were not among the provisions blocked in the orders we reviewed.

What Texas is asking the court to do

The petition asks for temporary and permanent injunctions. An injunction is a court order requiring a party to do or stop doing something. The requested orders would bar OpenAI from:

  • Offering ChatGPT to minors in Texas.
  • Releasing any new AI model to Texas users unless an independent evaluator approved by the court has reviewed its safety guardrails and approved the release.
  • Operating a product that provides content unsuitable for children without age verification, or that promotes compulsive use.
  • Falsely representing ChatGPT's reliability, safety or human-like qualities.

Texas also asks for civil penalties of $10,000 for each DTPA violation, disgorgement of assets, interest, court costs and attorneys' fees. Disgorgement means giving up money gained through unlawful conduct.

Nothing has changed for ChatGPT users yet None of these orders is in effect. ChatGPT remains available in Texas, including to teens, unless a judge grants the state's requests.

Timeline

Date Event
December 2015 OpenAI founded as a nonprofit
September 1, 2024 Texas SCOPE Act takes effect
September 2025 OpenAI adds optional parental controls, according to the petition
August 2026 ChatGPT for Teens launches, according to the petition
October 7, 2026 Common Sense Media rates ChatGPT for Teens an "Unacceptable Risk," as cited in the petition
October 23, 2026 Texas files its petition in Collin County district court

What happens next

The defendants must be formally served and then respond. Companies sued by states often challenge jurisdiction or ask courts to dismiss some claims, and Texas has stated in the petition that it brings only state-law claims and disclaims any federal claim.

Because Texas asked for a temporary injunction, the court could hold an early hearing on whether to restrict ChatGPT in Texas while the case proceeds. No hearing date was public when we checked. The case would then move to discovery, the stage where each side gathers documents and testimony.

The suit joins other legal pressure on OpenAI over minors. A separate proposed class action accuses the company of collecting children's data without parental consent; see our coverage of the ChatGPT children's data lawsuit.

What remains unclear

  • The cause number. The published copy of the petition shows a blank cause number and judicial district.
  • Which allegations Texas can prove. Many factual claims rely on news reports, studies and other lawsuits rather than Texas investigation findings.
  • How the SCOPE Act challenges affect this case. Court rulings on the law's constitutionality could shape Count II.
  • Whether the requested limits are possible. Barring minors or requiring court approval for new models would be unusual remedies, and OpenAI is likely to contest them.
  • Political context. The Texas Tribune reported that Elon Musk, whose company xAI competes with OpenAI, is the largest donor to Paxton's Senate campaign. The petition does not mention this.

This article is for information only and is not legal advice. For cases consumers can act on now, see our list of class action cases you can join.

Frequently asked questions

Why is Texas suing OpenAI?

Texas says OpenAI and Sam Altman marketed ChatGPT as safe and reliable without warning users about its risks, and that the company fails to verify parents or give them the tools required by the Texas SCOPE Act.

Is this a class action I can join?

No. It is an enforcement lawsuit brought by the State of Texas. There is no class, claim form or deadline for individual users.

What is the Texas SCOPE Act?

It is a 2023 Texas law, effective September 1, 2024, that requires online services to limit data collection from known minors and give verified parents tools to manage a child's account. Parts of it have been blocked by a federal court.

Could ChatGPT be banned for minors in Texas?

Texas asks the court to bar OpenAI from offering ChatGPT to minors in the state. That would happen only if a judge grants the request, and OpenAI can oppose it.

How much money could OpenAI owe?

The DTPA allows civil penalties of up to $10,000 per violation. The petition does not estimate a total, and any penalty would be decided by the court or a jury.

What did OpenAI say about the lawsuit?

An OpenAI spokesperson told The Texas Tribune that the company has worked with experts to mitigate risks and is "committed to working with Texas and other states on advancing pragmatic AI policies."