Texas Sues OpenAI With Florida’s Law Firm and Wording as Paxton’s Senate Race Closes

Texas sues OpenAI and its chief executive, Sam Altman, in a 67-page petition filed in a district court in Collin County on Friday 23 October, four days after early voting opened in the US Senate race that the state’s attorney general, Ken Paxton, is trying to win. The case accuses OpenAI of deceiving Texans about ChatGPT’s safety and reliability and of breaking a state law on children’s online safety.

Its central passages repeat, almost word for word, the complaint Florida filed in June, and the same private law firm works on both cases. AI rules in the United States are increasingly written in lawsuits brought by elected prosecutors, and in Texas the lawsuit arrived while the prosecutor bringing it is on the ballot.

What Texas wants from OpenAI

Paxton’s office relies on two statutes: the Texas Deceptive Trade Practices-Consumer Protection Act (DTPA) and the Securing Children Online through Parental Empowerment (SCOPE) Act, a children’s online safety law passed by the Texas legislature in 2023. The petition names six OpenAI entities alongside Altman, whom it seeks to hold personally liable for statements he made about his company’s products. It asks for civil penalties of $10,000 per violation and classes the monetary relief it seeks as above $1 million.

The requested injunction would reach into OpenAI’s product decisions. Texas wants the court to bar minors from ChatGPT in the state, restrict features that encourage compulsive use, and require any new OpenAI model to pass review by a court-approved independent evaluator before it reaches Texan users. The voluntary federal programme for reviewing models before release cannot hold back or stop a launch, so Texas is asking a state judge to make pre-release review compulsory.

OpenAI said governments have a role in setting strong safety standards and that it wants to work with state governments, Texas included, on rules for every AI developer rather than one company alone. Paxton said in the attorney general’s announcement that OpenAI is responsible for protecting its users and alerting them to the risks of ChatGPT, and that he intends to hold large technology companies to account under Texas law.

The same harm list and the same law firm

Florida’s attorney general, James Uthmeier, sued OpenAI on 1 June. The fifth paragraph of the Florida complaint lists mass shooters aided in their attacks, vulnerable people pushed towards suicide, professionals publicly humiliated, users losing critical thinking skills and minors hooked on a tool that imitates compassion to collect their data. The Texas filing carries the same list and blames it on the same motive, OpenAI’s pursuit of AI dominance and personal wealth while aware of the dangers.

Keller Postman, a plaintiffs’ firm, is outside counsel to Florida in its OpenAI case and to Texas in this one, where Ashley Keller works alongside Tony Buzbee, the lawyer who defended Paxton at his 2023 impeachment trial. Paxton has used the firm before. Texas settled its privacy case against Meta for $1.4 billion in 2024, and Keller Postman billed the state $97 million for that work under a contingency arrangement, according to the Texas Tribune.

Hiring outside counsel lets an attorney general’s office take on companies with far larger legal budgets. It also lets one legal theory travel between states with little rewriting, and Florida’s version already rests on a tobacco-style liability argument that treats ChatGPT as a product its maker knew to be dangerous.

A lawsuit filed inside a Senate race

Paxton is the Republican nominee for the US Senate and faces Democrat James Talarico on 3 November, with early voting running until 30 October. Polls published last week ranged from a tie to a four-point lead for Talarico. Elon Musk, the largest known donor of this election cycle, gave $158 million in September to the Senate Leadership Fund, which passed matching sums to Texas PAC, a super PAC backing Paxton, according to federal filings reported by several outlets. Crypto and technology money has entered other Senate contests this cycle too, including Fairshake’s $30 million commitment against Sherrod Brown in Ohio.

Musk’s AI business makes the Grok chatbot, a direct competitor of ChatGPT, and he lost his own case against OpenAI in May, when a federal jury found his claims time-barred after less than two hours. The Texas Tribune pointed out that Texas has not acted against Musk over sexual deepfakes made with Grok, a matter California’s attorney general is investigating.

The legal theory did not start in Texas and does not belong to one party. Florida’s Republican attorney general filed first, and on 14 October eight states with attorneys general from both parties, including Arizona’s Democrat Kris Mayes, filed coordinated consumer protection suits against OpenAI in their own courts. Texan voters already judge the candidates on the subject: in a Texas Public Opinion Research poll of likely voters, 47% said Talarico would better handle AI and data centres, against 42% for Paxton.

Why the states move while Washington waits

The White House has spent ten months trying to stop states from regulating AI. Executive Order 14365, signed by President Donald Trump on 11 December 2025, set up a Justice Department task force to challenge state AI laws and asked his advisers to draft legislation pre-empting them. The order’s own text told those advisers not to propose pre-empting state child safety protections. Neither statute Texas used was written for AI, and the pre-emption campaign was built around laws that were, such as Colorado’s AI Act.

Congress has not filled the space either. Senators stripped a ten-year moratorium on state AI laws from the 2025 budget bill by 99 votes to one, and Altman declined to appear at a September Senate hearing where Senator Josh Hawley called for laws making AI developers liable for large-scale failures. Ten state attorneys general, each answering to their own voters, are now pressing their own terms on OpenAI through the courts instead.

What the case reaches beyond Texas

The petition begins its argument by noting that AI now represents more than a third of US market value, and claims those gains came from deceiving and exploiting users. It also states that OpenAI collects Texans’ information about their health, finances and relationships through their conversations with ChatGPT. Fintech companies building on these models face the same logic, in which data handling becomes a consumer protection question a state can take to court.

Chatbots are moving closer to money. ChatGPT added a personal finance feature in May that connects US Pro users’ bank accounts through Plaid, and the accuracy claims Texas calls deceptive are the same claims that make such features worth using. The petition cites bad medical and legal advice and lists the public humiliation of professionals among its harms, which places lawyers, consultants and analysts among the consumers Texas says were wronged.

AI labs have made safety part of how they compete for users and investors, a safety-as-marketing strategy that now has a legal cost. The Texas and Florida theory turns each public promise of safety, accuracy or reliability into a statement a court can test against what the product did.

The rules for AI products in the United States are now being proposed in state consumer protection complaints, drafted by elected prosecutors and the private firms they hire, in the child safety lane that the White House’s own pre-emption order left to the states.

Editor’s note

Every piece goes through careful verification, but mistakes can happen. Readers who spot an error or have additional information can write to rosalia@thebrightminded.com.