Texas Sues OpenAI, Seeks Court Sign-Off on New Models

🎙️ Listen to this post: Texas Sues OpenAI Over ChatGPT, Demanding Court Sign-Off on Future Models

0:00 / --:--

Ready to play

Abstract editorial illustration of a courthouse column and an AI chat-bubble motif balanced on a scale, symbolizing a state lawsuit against an AI company

Last updated: 26 October 2026. Figures below are sourced to primary court filings, state attorney-general offices and news agencies linked inline; any number reported by only a single outlet is flagged as such in the text.

Contents

🎙️ Listen to this post: Texas Sues OpenAI Over ChatGPT, Demanding Court Sign-Off on Future Models The 60-second version Key dates and numbers What Texas actually filed Why this is the tenth suit, not the first The SCOPE Act gives Texas a weapon other states lack Litigation is doing the job regulators haven’t What this means if you build or publish with AI What we still don’t know FAQ Sources

The 60-second version

On 23 October 2026, Texas Attorney General Ken Paxton sued OpenAI and Sam Altman personally in a Collin County district court, alleging the company deceived users about ChatGPT’s safety and failed to protect minors.
Texas is the tenth US state to sue OpenAI over ChatGPT’s effect on children this year, following Florida in June and a coordinated filing by eight states — Arizona, Arkansas, Idaho, Iowa, Kansas, Montana, Nebraska and West Virginia — on 14 October.
Texas’s suit stands out for one specific demand: a court-appointed, independent evaluator that would have to approve any new OpenAI model before it reaches Texas users.
It leans on a law most other states don’t have — the SCOPE Act, Texas’s 2023 child-data statute — alongside the state’s Deceptive Trade Practices Act.
OpenAI has not filed a formal court response; a company spokesperson said governments have “an important role to play in setting robust safety standards for AI” and that OpenAI wants to work with Texas and other states on policies for “the entire AI industry — not just one company.”

Key dates and numbers

Date	Event
1 June 2026	Florida files the first state lawsuit against OpenAI over ChatGPT and minors.
5 October 2026	FTC opens a federal safety probe into OpenAI and Anthropic.
14 October 2026	Eight states (AZ, AR, ID, IA, KS, MT, NE, WV) file consumer-protection suits against OpenAI in their own state courts.
23 October 2026	Texas sues OpenAI and Sam Altman in Collin County, citing the DTPA and the SCOPE Act.
$5,000–$50,000	Range of per-violation civil penalties sought across the various state suits, depending on the statute invoked.
>$1 million	Monetary relief Texas is seeking, stated directly in the filed petition’s jurisdictional allegations.

What Texas actually filed

The petition, filed in a Collin County district court on 23 October, names several OpenAI corporate entities (OpenAI Global, OpenAI Foundation, OpenAI OpCo, OpenAI Group PBC, OpenAI Holdings and OpenAI, LLC) and Sam Altman personally as defendants, and accuses the company of violating the Texas Deceptive Trade Practices–Consumer Protection Act by marketing ChatGPT as safe, accurate and reliable while allegedly failing to warn users about its risks. A second count invokes the Securing Children Online through Parental Empowerment (SCOPE) Act, a 2023 Texas law requiring digital services likely to be used by minors to verify ages and give parents tools to manage a child’s data and advertising exposure.

According to WBAP/Dallas Express and The Texan, the petition opens with a 2015 Sam Altman quote — “AI will probably most likely lead to the end of the world, but in the meantime, there’ll be great companies” — before arguing that AI overall “now accounts for over a third of U.S. market value” and that OpenAI’s own rise to the center of that boom “is attributable to a web of deceit and the exploitation of users.” Its description of harm, confirmed directly against the filed petition, states that “mass shooters have been aided and abetted in deadly rampages, vulnerable people have been encouraged into suicide, professionals have suffered public humiliation, users have lost critical thinking skills, and minors have become addicted to a tool that feigns human compassion to collect their data with no parental oversight,” as The Texas Tribune also reported.

“I will fight to protect Texans from the harms of AI and ensure that powerful technology companies are not above the law,” Paxton said in a statement announcing the suit, per The Texan and WBAP. The filing itself is blunter about motive, stating that the alleged harms are “driven by the company’s insatiable quest to win the AI arms race and amass large fortunes, despite knowing the dangers of ChatGPT.”

- Advertisement -

Texas wants an injunction blocking ChatGPT access for minors in the state and requiring an independent, court-appointed evaluator to review safety protections on any future OpenAI model before it reaches Texas users — a condition no other state suit has requested. The petition itself confirms Texas is seeking monetary relief “over $1,000,000” in penalties and other relief; WBAP additionally reports a $10,000-per-violation civil-penalty figure that we could not verify against the petition’s prayer for relief, so treat that specific number as unconfirmed.

Why this is the tenth suit, not the first

Texas’s filing looks dramatic in isolation, but it lands at the end of a five-month build-up. Florida’s attorney general filed the first state suit in June. Then, on 14 October, eight more states — Arizona, Arkansas, Idaho, Iowa, Kansas, Montana, Nebraska and West Virginia — filed separate but loosely coordinated consumer-protection suits, several announced together at a joint press conference, according to OpenClassActions’ review of the filed complaints. Those suits lean on a patchwork of statutes — Arkansas’s children’s online privacy law, Montana’s Consumer Data Privacy Act, Nebraska’s Age-Appropriate Online Design Code — with civil penalties ranging from $5,000 per violation in Idaho and West Virginia to $50,000 per violation under Nebraska’s design-code statute.

What makes Texas different isn’t the underlying complaint — deceptive safety marketing, weak age verification, thin parental controls — but the remedy it wants. No other state suit asks a court to install a gatekeeper over OpenAI’s release schedule. Granted even narrowly, that request would hand a Texas court, not OpenAI’s safety team or the FTC, a veto over whether a given model ships to Texas users at all. That’s a materially bigger ask than damages or an age gate, and it’s the detail most headlines buried under the word “lawsuit.”

Paxton has also used the DTPA against Meta, TikTok, Snapchat and Roblox recently, per the Texas Tribune, and opened a 2025 investigation into Meta AI Studio and Character.AI over alleged deceptive mental-health claims. Texas’s approach to AI companies is a pattern, not a one-off.

The SCOPE Act gives Texas a weapon other states lack

Most of the other nine state suits argue, in effect, that ChatGPT was unsafe and OpenAI said otherwise. Texas makes that argument too, but the SCOPE Act count gives it a narrower second one: that ChatGPT is a “digital service likely to be used by a minor” that skipped the specific, enumerated things the statute requires — age verification, parental account controls, no targeted ads to minors without consent. That’s a lower bar than proving ChatGPT caused a specific tragedy; it only requires showing the compliance gap existed.

- Advertisement -

The SCOPE Act’s record in court is mixed, not the clean win for Texas that some coverage suggests. NetChoice and the Computer & Communications Industry Association sued to block the law in 2024; a federal judge issued a preliminary injunction against its content-monitoring and filtering requirement, and the Fifth Circuit affirmed that injunction this year on the grounds that the provision is preempted by Section 230 — while a separate requirement that platforms with public profiles verify users’ ages was left in place, simply because it wasn’t part of that injunction. No appeals court has actually ruled on the SCOPE Act’s age-verification requirement under the First Amendment. (A different Texas law — the one requiring age verification on adult-content sites, which the Fifth Circuit and the Supreme Court upheld in 2025 — is sometimes confused with the SCOPE Act; they are separate statutes.) What that leaves OpenAI is still a narrower set of defenses than a negligence claim would face: Texas only has to show the compliance gap existed, not prove ChatGPT caused a specific harm. OpenAI’s account-linking parental controls from September 2025 and its “ChatGPT for Teens” feature from August 2026 — both described in the petition itself — will likely be the centrepiece of its defense; the petition argues the voluntary, opt-in nature of that linking is insufficient.

Litigation is doing the job regulators haven’t

Set against this month’s other AI-policy news, the pattern is clear: enforcement against frontier AI companies in the US is happening in state courtrooms, not Washington. The FTC’s safety inquiry into OpenAI and Anthropic, opened 5 October, is still an inquiry, not a complaint, and the administration has otherwise favoured voluntary accords over binding rules. Brussels has moved further: the EU’s AI Office opened its first AI Act probe into OpenAI and Anthropic on 16 October, and the UK’s ICO separately secured privacy changes from ten AI developers without going to court at all. Ten US state lawsuits are what fills that vacuum when Congress hasn’t passed AI-specific legislation and the federal regulator with the clearest mandate is still gathering information.

This is also happening against a company whose recent record gives plaintiffs’ lawyers plenty of material. OpenAI disclosed in September that it had shelved a planned model release over safety failures while shipping autonomous “Dots” agents. None of that is cited directly in the Texas petition as far as we can confirm, but it’s the backdrop against which “the company knew and shipped anyway” arguments tend to land well with a judge or jury — and it makes the $1.4 trillion valuation behind OpenAI’s recent $30 billion funding round a tempting target for state civil-penalty math running at $10,000 to $50,000 a violation.

- Advertisement -

What this means if you build or publish with AI

For teams shipping consumer-facing AI products, four things here are worth acting on now rather than waiting for a verdict:

Age verification is no longer a theoretical compliance risk. If your product is “likely to be used by a minor” under a state’s children’s-online-safety statute, voluntary, opt-in parental controls may not satisfy laws modelled on the SCOPE Act. Mandatory, verifiable age gates are the direction several of these suits push toward.
Marketing copy about “safety” and “reliability” is now litigation evidence. Every one of the ten state suits leans on consumer-protection law, which turns product marketing into a liability surface.
Expect a state-by-state patchwork, not one US standard. Nebraska’s design code, Montana’s data-privacy law and Texas’s SCOPE Act impose different, overlapping obligations; a product compliant in one state may not be in another.
Document your model-release safety process now. If courts start granting relief resembling Texas’s request — third-party sign-off before release — companies with an already-documented internal evaluation process will be far better placed than those improvising one under a court order.

What we still don’t know

We read most of the filed petition directly against the Texas Attorney General’s own copy, but could not retrieve its prayer-for-relief section; the exact per-violation penalty figure Texas is requesting (reported by WBAP and others as $10,000 per violation) remains unconfirmed against the document itself.
Whether Texas coordinated timing or strategy with the earlier eight-state filing, or acted independently, isn’t established in any source we reviewed.
How a Texas court would actually implement a “court-approved evaluator” for model releases — who would staff it, what standard it would apply — is unaddressed in any reporting so far.
OpenAI has not filed a substantive legal response as of this writing, so its specific defenses to the SCOPE Act claim aren’t yet on the record.
Whether the FTC’s separate federal inquiry will produce any formal action that overlaps with, reinforces, or is overtaken by these state cases remains unclear.

FAQ

Is ChatGPT currently banned for minors in Texas?
No. The ban is a remedy Texas is asking a court to order, not a law already in effect.

How is Texas’s suit different from the eight states that sued on 14 October?
Same core allegations, but Texas alone asks a court to require independent, court-approved review of future OpenAI models before Texas release, and it alone invokes the SCOPE Act.

What is the SCOPE Act?
Texas’s 2023 child-safety law requiring digital services likely used by minors to verify ages and give parents control over a child’s data and ad targeting. Courts have blocked its content-filtering requirement as preempted by federal law but left its age-registration provision in place; no appeals court has yet ruled on its age-verification requirement under the First Amendment.

Has OpenAI responded?
Not with a formal court filing yet. A company spokesperson said governments have “an important role to play in setting robust safety standards for AI” and that OpenAI is committed to working with Texas and other states on policies for “the entire AI industry — not just one company,” without disputing the specific allegations in detail.