xAI sued in Arkansas for using Grok to create deepfake child pornography of real minors

By Zendoric · July 25, 2026.

An Arkansas family has sued xAI, Elon Musk's artificial intelligence company, after police confirmed that their daughter's image was used to generate child sexual abuse material (CSAM) with the Grok model. The lawsuit, filed by Potts Law Firm and reviewed by KATV, is not an isolated case: according to its managing partner, Derek Potts, the firm is already preparing additional lawsuits on behalf of other families —adults and minors— who have been notified by police that their image met the same fate, and it will file them next week.

The case originates from a criminal investigation. Russell Bloodworth III, a photographer known in northwest Arkansas, was arrested on June 10 and charged with 200 counts of possessing and creating AI-generated child sexual abuse material. According to Bentonville police, Bloodworth used Grok to transform photographs he had himself taken of underage clients into sexually explicit content; a search of his phone, documented in the police report, revealed nearly 1,700 images and videos of this kind.

The lawsuit against xAI seeks financial damages, the deletion of all illegally generated material and, above all, that the company implement safeguards to prevent Grok from generating nonconsensual sexual depictions of real people, including minors. Potts is explicit about why he is targeting the model and not just the user: "I think everyone agrees that this platform, compared to others, is unique, that it marketed itself as 'rebellious.' They promoted additional modes to create what they call 'spicy' images," he told KATV, in reference to the image-generation feature without the usual restrictions that xAI has used as a selling point for Grok against more conservative rivals.

The regulatory vacuum is the second pillar of the case. Arkansas has passed laws criminalizing the creation of CSAM via AI, but, as Potts notes, "there are virtually no laws passed by either Arkansas or the federal government to restrict, limit or regulate AI itself." Phillip Huff, a computer science professor at the University of Arkansas at Little Rock, frames it in broader terms: "I don't think we can continue living in a world where only a few companies set all the rules on AI and we have to live with the consequences."

This is exactly the kind of case where our usual thesis on regulation —beware of legislating out of panic rather than out of evidence— does not apply in reverse: here the evidence already exists, it is a police report with 1,700 files and 200 counts, not a speculative projection about future risks. When a company chooses to differentiate itself commercially by offering less friction than its competitors in generating images of real people, the absence of technical barriers ceases to be a neutral product decision and becomes, potentially, a factor of civil liability. It is plausible that this litigation, and those the firm announces for next week, will become the real vehicle for establishing what obligations generative-model providers have when their design facilitates —though does not mandate— a crime with identifiable minor victims.

It is worth separating two levels, because they are easy to confuse. One is the long-term outlook for AI that we defend at Zendoric: increasingly capable models freeing up time and resources, with the potential to transform medicine and material abundance. The other is this specific case, where the design of a product —explicitly marketed as less restricted— has facilitated, according to the police charges, the victimization of real minors. There is no possible nuance here: protecting children from AI-generated abuse is not a "transition cost" that resolves itself over time, it is a red line that demands technical barriers and legal accountability now, regardless of how optimistic the underlying reading of where this technology is headed may be. If anything, this case confirms that the near-total absence of specific federal regulation on generative AI models —beyond scattered and reactive state laws, like Arkansas's own— leaves the bar for prevention in the hands of the courts, lawsuit by lawsuit, instead of in clear rules set in advance.