Texas has made ChatGPT a consumer-protection defendant. Attorney General Ken Paxton sued OpenAI Global, LLC, the OpenAI Foundation, four affiliated entities and CEO Sam Altman personally in Collin County District Court on Friday, October 24 (US time), according to Courthouse News, alleging the company misled Texas consumers about ChatGPT’s safety and failed to implement sufficient safeguards — claims framed under the Texas Deceptive Trade Practices Act and the state’s SCOPE Act, which covers minors’ digital services. OpenAI has not yet responded publicly; none of the allegations has been tested in court.
🔍 THE BOTTOM LINE: The reason this filing matters is not the count — attorney-general suits against OpenAI have run all year — it is the recipe. The Texas filing combines three things no earlier suit has assembled in one petition: a consumer-deception claim built on the DTPA, a minors’ statute (SCOPE) violation, and personal liability for Altman as the person who “personally directed” ChatGPT’s design and safety policies. The injunction ask takes the same maximalist turn — new ChatGPT models blocked in Texas until a court-appointed evaluator approves them. If the recipe catches on, every US state with a DTPA analogue now has a template.
What the complaint actually alleges
The 67-page petition files its core under §17.47 of the DTPA — the state’s deceptive-trade-practices enforcement mechanism — plus the SCOPE Act’s §509.151, which expressly routes platform violations into DTPA enforcement. The state’s Consumer Protection Division chief Jonathan Stone writes that ChatGPT “is designed to act like a friend to encourage use of the chatbot and then betrays any vulnerable adolescents to whatever frightening need they bring to their ‘confidant’”, calling the design feature, in the state’s framing, “at very best… reckless, and at worst… willful”. The petition cites at least eight instances of people who committed violent acts after consulting ChatGPT, as Courthouse News summarised — including the 2025 Florida State University shooting, whose family’s separate suit claims the shooter asked ChatGPT “how many people he’d need to kill to become notorious”.
The relief sought is where this gets unusual: monetary penalties under the DTPA, court orders requiring changes to ChatGPT’s practices, disgorgement tied to the SCOPE Act claims, personal liability for Altman — and a standing injunction that would block OpenAI from releasing new ChatGPT models in Texas without approval from a court-appointed independent evaluator, per Courthouse News’s reading of the filing. Texas also invokes the company’s own original mission language — “unconstrained by a need to generate financial return” — as the baseline it says OpenAI abandoned.
The political frame is in the filing
The suit lands with Paxton mid-campaign for the US Senate seat he is challenging for, a week after filings showed Elon Musk’s super PAC spent nearly $200 million backing his bid, as Courthouse News noted. The filing itself quotes Musk’s “end of the world… but in the meantime, there’ll be great companies” line from a 2015 Future of Life Institute post. Both sides of the context matter for reading it: the enforcement is timed to a campaign, and the underlying legal architecture — DTPA plus SCOPE — exists independently of who filed it, with the state’s press release citing SCOPE investigations that predate the campaign window.
The Trump administration’s position sharpens the tension. Washington has pushed a voluntary, deregulatory federal posture while states — Republican and Democratic attorneys general alike — keep building the enforcement record the federal government declines to write. Paxton’s petition joins the eight-states coalition two weeks ago and Florida’s first-in-nation suit in June, each layer adding a new legal theory: Florida went after Altman’s personal assets, the coalition built the coordinated front, Texas adds the offline-release injunction and a state minors’ statute.
The industry picture: a compliance surface being built one state at a time
For the product itself, the accumulating litigation is designing ChatGPT’s future safety surface in public, whether or not any single case wins. A court-approved evaluator for new model releases, age verification parents can verify, SCOPE-grade parental oversight — these asks will not all survive contact with the Texas judiciary, but some fraction of them will standardise. The historical pattern from tobacco and opioids, which several AGs have invoked by name, is that consumer-protection settlements become de facto design standards long before legislatures act. OpenAI’s most recent public posture remains the one from August: industry-leading age-specific protections, parental controls, break reminders, and a preference for “pragmatic AI policies that apply to the entire AI industry, not just one company.”
For New Zealand the signal is about whose defaults we inherit. As we wrote when the eight-state coalition filed, compliance architecture from US litigation tends to become the global product default — the same chatbot a Tauranga teenager opens next month is being redesigned right now in a Collin County courtroom. And while a federal NZ AI bill remains unwritten, our own regulator has already set frontier-AI cyber expectations — meaning the Texas petition’s demand for independent evaluators is no longer only an American debate. The petition also cites the July Hugging Face breach, which we covered — safety failures now travel into consumer-deception claims as evidence, not just into specialist AI-safety coverage.
The practical detail to watch next: OpenAI’s response deadline, and whether Texas courts engage with the evaluator-injunction ask on the merits or treat it as out of bounds for a product-safety claim. Either answer writes precedent for what a state can demand of a frontier lab.
Sources: Courthouse News, Just the News, Texas OAG, Texas Scorecard
