xAI, owner of the Grok chatbot and image generator, is suing the state of Minnesota to defend its right to produce and host nonconsensual AI-generated nudified images.
Of course, that’s not exactly how xAI words it. Instead, they present themselves as a warrior for the First Amendment, keeping the citizens of Minnesota safe from a censorious state government that doesn’t want to have a little AI fun.
The lawsuit stems from a first-of-its-kind nudification ban meant to focus on the technology that makes nudification possible. While most of the focus in legislating AI nudification has been on on after-the-fact punishment of those creating nudified images, Minnesota lawmakers sought to cut off the ability and accessibility to create nudified images at the source. They wanted to hold AI companies responsible because while a user might prompt the AI to produce a certain image, it’s ultimately the AI — not the user — doing the work.
The lawsuit stems from a first-of-its-kind nudification ban meant to focus on the technology that makes nudification possible.
In cloud-based services like Grok, AI-generated images are created by the AI company using company-owned hardware in company-owned data centers. When users take a photo on X and tell Grok — and these are real examples — “remove pants and shirt” or “spread her legs” it is xAI servers receiving that request and xAI graphics processors generating a response. The user has put no skill or effort into this digital sexual abuse but has reaped the illicit benefit from a company that has the skill to digitally alter such an image. Thus, Minnesota argues, the tech companies doing the actual image generation are as much to blame as their users making such requests.
xAI, of course, sees things differently. They argue the law is “an overbroad, content-based ban on free speech and the tools of visual expression.” While they tacitly admit the ban on nudification would “implicate Grok Imagine” and that they have an “actual and well-founded fear” that the law would be used against them, their legal argument is rather cagey about whether or not their AI is all-too-eager to nudify.
Which we know it is.
The context
I’ve written a couple of times now about Grok, the purposefully edgy chatbot with a rebellious streak. In late 2025, Grok was integrated into the X/Twitter platform so anyone who called on the name of @grok could summon the AI to do its bidding — even if that bidding was to digitally nudify images of unsuspecting and nonconsenting people.
The immediate outcome was entirely predictable. X users began nudifying and sexualizing images at will. Despite such immediate and obvious harm, it took 11 days — and pressure by several national governments — for xAI to remove the imaging editing feature from X. During those 11 days, researchers estimate Grok generated more than 3 million sexualized images — about 65% of all images it produced.
After a week and a half of outage, xAI shut down public access to their digital sexual abuse generator. So they could monetize it.
Rather than realize the harm they had done, make reparations and tighten their guardrails to ensure nudification was not possible, xAI chose instead to limit Grok’s image editing and nudification capabilities to paid subscribers. According to SEC filings, on Dec. 31, 2025 — in the middle of the 11-day free-for-all — Grok had 4.9 million active paid subscribers. By March 31, 2026, that number had surged to 6.3 million.
(Shutterstock)
Other analyses show on the day of the ban, mobile in-app purchases on X skyrocketed by 18% while, in early January, downloads of the Grok app went up 72% compared to the previous year.
Now, correlation is not causation, but it’s not too huge a leap to surmise that this so-called spicy mode was xAI founder Elon Musk’s on-brand method of popularizing his AI platform — which is, in turn, a subsidiary of SpaceX — in the build-up to SpaceX’s record-breaking initial public offering which made Musk the world’s first trillionaire.
While the furor has died down and xAI has since implemented some guardrails to prevent nudification, the fact remains that Grok Imagine has the ability to take a normal photo of someone, listen to a user’s illicit imaginings, and bring that imagining to digital reality.
In the court filing, xAI argues they should not be responsible for how their image generator responds to user prompts, complaining to the court that Minnesota’s law has “no safe harbor for good-faith efforts of the provider of general-purpose AI creative tools to avoid harms.”
In other words, “We can’t stop our Elon Musk-coded chatbot from creating sexual abuse imagery but that’s not our fault.”
Minnesota believes that it is their fault.
Minnesota’s nudification ban
Minnesota’s path to banning nudification technology began in February 2026 when three women testified before the Minnesota House Commerce Finance and Policy Committee about how their private Facebook photos had been nudified and transformed into explicit photo-realistic AI-generated pornography. In response to this testimony, Rep. Jessica Hanson drafted Minnesota HF-1606.
The law doesn’t single out Grok or xAI specifically but explicitly prohibits all websites, apps, software and other services from allowing users to access, download or use their systems to alter or generate a realistic image or video of a real, identifiable person to depict “intimate parts” — defined as genitals, groin, inner thigh, buttocks or breasts — that were not present in the original photo.
The penalty for violating the law is a fine of up to $500,000 per instance with all collected fines being used to fund direct services and advocacy for victims of sexual assault, general crime, domestic violence and child abuse. The law also states those violating the law may be subject to civil cases by victims to seek compensation for damages.
It went to a vote in the Minnesota House of Representative April 23 and passed in overwhelmingly bipartisan fashion, 132-1. Six days later, it went to the Minnesota Senate, where it passed unanimously, 65-0. Gove. Tim Walz signed the bill into law May 7, with implementation to begin Aug. 1.
xAI’s lawsuit
Four days before the law was to go into effect, xAI filed a lawsuit to stop it, seeking an immediate preliminary injunction to keep the law from going into effect until the law could be argued on its constitutional merits.
Four days before the law was to go into effect, xAI filed a lawsuit to stop it.
On Sept. 4, the district court denied that injunction, not ruling on the law’s constitutionality but declaring that xAI had not established that the law’s implementation would irreparably harm the $250 billion corporation. The court emphasized the legislature’s near-unanimous enactment of the law, the state’s undeniable public interest in protecting its citizens, and the undisputed harms from AI nudification.
xAI then appealed to both the U.S. Department of Justice and the U.S. Eighth Circuit Court of Appeals. The Trump DOJ sided with xAI, filing a statement of the interest in the case, telling the court they believe the federal TAKE IT DOWN Act is sufficient.
Notably, that law — bipartisanly co-sponsored by Sens. Amy Klobuchar and Ted Cruz — does not penalize AI companies for their role in sexual abuse imagery creation. The Trump administration is worried that imposing such penalties “risks paralyzing the AI industry and hampering United States leadership in AI.” Leadership that, apparently, includes bombing schoolchildren, making propaganda and creating racist video games.
Meanwhile, the conservative-dominated Eighth Circuit Court of Appeals overturned the distinct court’s denial and granted the injunction. In a brief, one-sentence order that did not explain its reasoning, the court temporarily paused any potential enforcement of the law until the case can be heard in full at the district level. There is currently no trial date set.
Who is to blame?
“Minnesota had the guts to go to the source. That’s why they’re receiving the pushback they are.”
In the end, this case is not really about nudification. It’s about who is responsible for it. If the Minnesota law didn’t target AI corporations but focused only on individual users, xAI would not be filing a lawsuit. Minnesota had the guts to go to the source. That’s why they’re receiving the pushback they are.
The question Minnesota has put before the courts is a simple one: When a company builds, owns, operates and profits from a machine capable of turning someone’s ordinary photograph into sexual abuse imagery, can that company then claim no responsibility for what the machine does?
Minnesota’s law represents an attempt to place responsibility somewhere upstream, where the technology and the people profiting from it actually exist. Whether the Constitution permits the state to do that is now for the courts to decide — and I’ll be frank, I’m no legal scholar but I’m not at all confident this will go Minnesota’s way.
But the underlying question should not be obscured by talk of constitutionality or AI freedom. When your product can be used to sexually abuse someone at the push of a button, how much responsibility are you willing to accept for what your button does? xAI says not much at all.
Josh Olds
Josh Olds is a public theologian and pastor for those disillusioned with institutional church. He is the creator of the small-group video series “Year on the Mountaintop” and a featured contributor to Fostering Hope: A Prayerbook for Fostering and Adoptive Parents. Follow his work on Facebook or at JoshOlds.com.