Texas wants a court-appointed expert to approve every new OpenAI model before release

The lawsuit opens with a 2015 quote from Sam Altman: AI will "most likely lead to the end of the world, but in the meantime, there will be great companies." With that quote, Ken Paxton, Texas's attorney general, goes after OpenAI and its CEO. The suit was filed on October 23 in Collin County court, the Texas Tribune reports.

He accuses the company of deceiving and harming Texans by presenting ChatGPT as safe, and of failing to put sufficient safeguards around it. Above all, he is seeking a sweeping injunction: OpenAI would have to block minors' access to its product and obtain approval from a court-appointed evaluator before releasing any new model.

What is Texas accusing ChatGPT of?

The attorney general's press release, cited by the Texas Tribune, lists a series of harms. Mass shooters allegedly received help and vulnerable people were allegedly pushed toward suicide. Professionals allegedly were humiliated in public, users allegedly lost their critical thinking, and minors allegedly became dependent with no parental control. According to the attorney general, OpenAI knew about these dangers and ignored them to dominate the AI race. The suit also targets Sam Altman personally.

Texas is not alone. Florida sued OpenAI and its CEO as early as June, and eight states each filed suit on October 14, accusing the company of misleading parents about ChatGPT's safety and harming children. The Texas request goes further: having every new model validated by a third party appointed by the court.

OpenAI's response

An OpenAI spokesperson, cited by the Texas Tribune, says the company has worked with experts to reduce risks and add safeguards. The spokesperson adds that governments have an important role to play in AI safety standards. OpenAI says it is ready to work with Texas and other states on rules that would apply to the whole industry, "not to a single company."

An attorney general, a competitor and a donor

That last remark takes on particular significance given the context detailed by the Texas Tribune. Elon Musk co-founded OpenAI, left it in 2018 after a dispute with Sam Altman, then founded its competitor xAI. He is now the top donor to Ken Paxton's Senate campaign, with nearly $200 million. xAI's chatbot, Grok, has been criticized for allowing the creation of fake pornographic images, including of children. California has opened an investigation into the matter. Texas, for its part, has taken no action against Elon Musk.

This is not Ken Paxton's first such move: over the past year, he has sued Meta, TikTok, Snapchat and Roblox, or reached settlements with them, in the name of consumer protection and especially child protection. For this case, the state hired outside lawyers, including Tony Buzbee, who defended him in his 2023 impeachment proceedings.

What to watch

A lawsuit is not a ruling. It will be up to the Collin County court to decide whether prior review of every new model can be imposed on OpenAI. In Europe, the AI Act works through legislation rather than the courts: it already sets obligations for providers of large models. The next step will be OpenAI's written response to the court.