Texas Asks a Court to Vet New OpenAI Models Before Release

Texas Attorney General Ken Paxton sued OpenAI and its chief executive Sam Altman on 23 October 2026, and among the orders his petition asks a court to make is one that would stop OpenAI from “Releasing any new artificial intelligence model to Texas users unless an independent third-party evaluator, approved by the Court, has reviewed its safety guardrails and approved its release.” The petition, read on 24 October 2026, was filed for the district court in Collin County and brings claims under the Texas Deceptive Trade Practices Act and the state’s SCOPE Act. These are allegations; no court has ruled on them.

Paxton’s office said he sued OpenAI “and its affiliated entities, including its CEO Sam Altman.” Its release quotes him: “With the rapid advancement of AI technology, OpenAI has the responsibility to protect users and warn them of the many dangers of its platform.”

Business Pill · GUARDRAILS

A short Business Engineer explainer on guardrails: rules enforced around an AI model, not inside it. It teaches the general idea only and says nothing about the lawsuit in this piece.

What the Petition Alleges

The petition opens with a line 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.” It then alleges that “Altman and OpenAI have failed to warn Texans of the dangers of their product, in violation of the Texas Deceptive Trade Practices-Consumer Protection Act.”

It has two counts. Count I is a claim under the Deceptive Trade Practices Act. Count II is a claim under the Securing Children Online through Parental Empowerment (SCOPE) Act, which the petition says requires “digital service providers,” like the defendants, to put in place specific safety and privacy measures for minors.

Altman is named in his own right. The petition says Texas “seeks to hold Altman personally liable for the harm he has caused Texans through his direct involvement in the conduct alleged as founder and CEO of OpenAI, and the deceptive statements he has personally made regarding his firms’ products.”

The petition also reaches for this week’s news. It alleges that “AI has begun to hack into customer accounts at several major national banks,” and its footnote for that line cites Axios’s 22 October report that unvetted AI agents are accessing U.S. bank accounts. Our bank-agents piece covered that report: Axios said the access usually comes through the front door, with the user’s permission and login credentials.

What Was Posted on X

Texas Attorney General Ken Paxton announced the suit on X on 23 October, saying he sued OpenAI and its CEO Sam Altman for lying to users and harming Texans. These are his allegations; the piece sets out what the petition itself asks a court to order.

What Texas Asks the Court For

The petition says Texas “seeks monetary relief over $1,000,000 and non-monetary relief.” The table lists the main requests in its prayer, in its own words.

What OpenAI Said

An OpenAI spokesperson said in a statement to The Hill: “As part of our focus on responsibly delivering the benefits of AI, we’ve worked closely with experts to continually improve ChatGPT, mitigate risks and build additional safeguards.” The statement continued: “Governments have an important role to play in setting robust safety standards for AI, and we’re committed to working with Texas and other states on advancing pragmatic AI policies that apply to the entire AI industry – not just one company.”

The Hill also reported that “Paxton is currently facing off against Democrat James Talarico in a tight race for the Lone Star state’s Senate seat.”

Connecting the Coverage

Texas adds to a run of state suits this month. Our eight-states piece reported on 14 October that Iowa’s petition asks for civil penalties of up to $40,000 for each violation of the Iowa Consumer Fraud Act, alongside suits from Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho and Montana. Per Iowa’s cap and Texas’s request, Iowa’s $40,000 per violation is four times Texas’s $10,000, by our arithmetic.

The same consumer-law route is being used against Meta. Our New Mexico piece reported that Meta must pay $25 billion in civil penalties in that state’s privacy case, under a law that allows up to $5,000 for each willful violation. Our Rhode Island piece reported that the state sued Meta on 22 October under its Deceptive Trade Practices Act over scam advertisements.

The Structural Read

As we read it, the petition asks for two kinds of remedy. One is money: $10,000 for each violation of the DTPA, disgorgement, fees and costs. The other is control over the product: limits on offering ChatGPT to minors in Texas, and a court-approved review before a new model reaches Texas users.

As we read it, naming Sam Altman as a defendant moves part of the case from the company to the person: the petition says Texas seeks to hold him personally liable for conduct it says he directed and statements it says he made.

As we read it, the petition also ties the case to this week’s agent news, citing Axios’s report on AI agents in bank accounts; that report described agents logging in with users’ own permission and credentials.

Three Implications

FOR AI LABS The petition asks a court to put an independent evaluator between a new model and Texas users; if granted, release timing in one state would depend on that review.

FOR EXECUTIVES The petition names the chief executive personally, alongside six OpenAI entities, and seeks to hold him personally liable for the conduct it alleges.

FOR THE STATES Texas uses its Deceptive Trade Practices Act and its SCOPE Act; Iowa’s petition used its Consumer Fraud Act and asks for up to $40,000 per violation, four times Texas’s $10,000 by our arithmetic.

The Business Engineer Lens

This story maps onto the Business Engineer framework The Product Overhang Doctrine.

The framework states: “The result is a permanent, widening gap, and that gap is now the most important strategic surface in software.”

As we read it, the gap the framework describes is between what models can do and what products ship. Texas’s request for a court-approved review before each new model reaches Texas users would add a step inside that gap. The framework was published in June 2026 and does not discuss this lawsuit.

What Is Not Established

No court has ruled on any of the petition’s claims, and the petition does not say how many violations Texas will allege, so no penalty total can be computed from it. The petition does not give a date for any hearing in the text we read. Whether a court would order the pre-release review Texas asks for is not established.

The Bottom Line

Texas’s petition asks for money, $10,000 for each violation of the DTPA, and for control over releases: no new OpenAI model for Texas users until an independent evaluator, approved by the court, has reviewed its safety guardrails and approved it. OpenAI said it is committed to working with Texas and other states on rules that apply to the whole industry.