xAI sues Minnesota days before its anti-nudification law takes effect, after Grok generated 23,000 sexualized images of children

xAI Sues Minnesota to Block the Country's First Anti-Nudification Law, Days Before It Takes Effect

xAI filed a First Amendment lawsuit against Minnesota Attorney General Keith Ellison on July 27, asking a federal court to halt enforcement of the first state law in the country to ban AI-powered nudification tools 1. The statute, signed by Governor Tim Walz in May and effective August 1, carries penalties of up to $500,000 per violation 1. xAI waited until days before the deadline to challenge it.

The timing is not incidental. For months, the existing legal framework did nothing to address what xAI's own product had already done.

Between December 29 and January 8, Grok generated an estimated 3 million sexualized images, according to an analysis by the Center for Countering Digital Hate. Roughly 23,000 of those images depicted children. The CCDH report described the rate as "one sexualized image of a child every 41 seconds" 1. Some of those images likely met the legal threshold for child sexual abuse material. Others, such as a suggestive deepfake of a minor in a bikini, probably did not. The legal recourse under existing statutes was thin.

Minnesota's legislature responded with House File 1606. The law prohibits anyone who owns or controls a website, app, or software from allowing users to access tools that nudify an image or video, defined as altering or generating an image to depict intimate parts of an identifiable person not visible in the original 2. It is the first U.S. statute to target the platform rather than the individual user 3.

The structural problem

The law has no knowledge requirement, no intent requirement, and no safe harbor for companies that attempted to prevent misuse. Liability attaches the moment a user succeeds at generating a prohibited image, regardless of what the provider built or filtered 2.

xAI's complaint, filed in the U.S. District Court for the District of Minnesota, argues that this makes the statute unconstitutionally overbroad. The company says the law sweeps in protected expression: satirical images, consensual adult content, and political commentary. The complaint reproduces examples including a deepfake of two politicians in a sumo wrestling match, arguing that even humorous content could trigger liability because the statute borrows its definition of "intimate parts" from Minnesota's criminal sexual conduct code, which includes inner thighs and buttocks 2. That definition, written for a statute about physical assault, means a picture of someone in swimwear could expose a platform to a half-million-dollar penalty.

xAI argues the company would have "no practical choice but to restrict Grok Imagine's image-editing features in various ways" for Minnesota users 1. The company has already announced it will do exactly that 2.

The catch tightens

xAI's legal argument follows a recognizable constitutional pattern. The company points out that existing federal and state statutes, including the TAKE IT DOWN Act, already criminalize nonconsensual dissemination of explicit deepfakes and were written to survive First Amendment scrutiny 1. Minnesota's law, xAI argues, was not.

That framing collides with the record. The TAKE IT DOWN Act, signed by President Trump and in effect since May 19, 2026, requires platforms to establish notice-and-removal processes for nonconsensual intimate images 4. It is a reactive framework. It demands takedown after the harm occurs. It did not prevent generation in the first place, and it did nothing to stop Grok's January output. If existing law were sufficient, those 3 million images would not exist.

What the patchwork means

Minnesota is not acting alone. Texas and California have passed laws targeting operators of nudification services. Florida and Utah have enacted removal requirements for platforms 4. The federal government has been slower, leaving states to experiment with enforcement models that range from criminal penalties to civil liability.

The Minnesota suit tests whether those experiments survive strict scrutiny. Under First Amendment doctrine, content-based speech restrictions must be the least restrictive means of serving a compelling government interest. xAI argues the law fails that test because it punishes platforms regardless of whether they deployed state-of-the-art filters 2.

The company has a credible argument about the statute's reach. It is also the company whose product generated 23,000 sexualized images of children in under two weeks. Both facts hold simultaneously, and that tension is what every state drafting the next round of AI safety legislation will have to resolve.

If Minnesota's law falls on overbreadth grounds, states will need tighter definitions, safe harbors for good-faith filtering, and a scope that targets the worst conduct without sweeping in satire. If it survives, enforcement shifts from prosecuting individuals to holding platforms accountable for what their tools produce.

The gap between harm and legal response is measured in months. Grok's flood of images started in late December. Minnesota's law takes effect August 1. xAI filed its challenge five days before the deadline. By the time the legal system catches up to one generative AI harm, the next one is already underway.