The xAI Minnesota nudification ban will remain enforceable while Elon Musk's AI company continues its constitutional challenge, after U.S. District Judge Donovan Frank rejected xAI's request for a preliminary injunction on September 4. The ruling leaves Minnesota's first-of-its-kind restriction on AI nudification tools in force during the litigation.

 

The decision is not a final judgment that the law is constitutional. Frank concluded that xAI had not shown the kind of injury needed to freeze enforcement at this stage, while acknowledging that the First Amendment questions are complex. The underlying lawsuit, X.AI LLC v. Ellison, remains active in federal court.

 

Minnesota's AI Nudification Law Stays Enforceable

Minnesota House File 1606 took effect on August 1 and created Minnesota Statutes section 325E.91. The law targets services that let users alter or generate images or video so that an identifiable person appears to have intimate body parts that were not shown in the original material.

 

Unlike many deepfake laws that focus mainly on the person who creates or distributes a harmful image, Minnesota's statute reaches the technology provider. It says a person controlling a website, application, software program or other service must not allow users to access, download or use that service to create prohibited nudified imagery.

 

The statute creates several direct obligations and remedies:

  • Providers may not allow users to access or use covered nudification functions.
  • Advertising or promoting services that perform prohibited nudification is barred.
  • People depicted in prohibited images can bring civil claims for damages and injunctions.
  • The attorney general can seek civil penalties of up to $500,000 for each unlawful access, download or use.

 

Minnesota legislative records also provide a limited exemption for services that require substantial technological or artistic skill from a human creator. The boundary of that exemption, and how it applies to general-purpose AI image editors rather than dedicated nudification apps, is one of the practical questions hanging over the law.

 

xAI's First Amendment Challenge Remains Open

xAI sued Minnesota Attorney General Keith Ellison in July, arguing that the statute is an overbroad, content-based restriction on protected expression. The company says the law can reach lawful or consensual image editing and imposes liability without a sufficiently protective safe harbor for platforms that try to prevent misuse.

 

The dispute reached Frank in two stages. He first denied xAI's emergency request to stop the law days before its August 1 effective date, pointing to the company's delay in seeking relief. He then considered the broader preliminary-injunction motion after an expedited hearing and again declined to block enforcement.

 

That procedural posture matters. A preliminary injunction asks a court to halt a law before the full case is decided, so the judge evaluates factors including likely success, irreparable harm and the public interest. Denying that request keeps the statute operating, but it does not resolve xAI's constitutional claims on the merits.

 

The state, meanwhile, argues that the measure is narrowly aimed at a documented category of harm: realistic sexualized images of identifiable people created without their consent. Minnesota officials have defended the law as a way to place responsibility on services that make those outputs possible, rather than relying only on victims to identify and pursue individual users after an image has already spread.

 

What the Ruling Means for Grok and Other Image Tools

The immediate compliance pressure falls on xAI because its Grok products include image-generation and editing capabilities. In earlier court filings, xAI said the threat of substantial penalties could force it to restrict features for Minnesota users while the case proceeds. The company has also argued that broad platform liability could affect general-purpose creative tools, not only apps marketed specifically for nudification.

 

For other AI developers, the case is a test of where states can draw the line between regulating harmful outputs and regulating the tools capable of producing them. A rule aimed at providers before an image is generated is structurally different from laws that punish distribution after the fact, because product design, geofencing, access controls and model safeguards become part of legal compliance.

 

The Minnesota statute therefore puts technical implementation at the center of a constitutional dispute. Providers may need to determine whether a feature can be disabled by jurisdiction, whether classifiers can reliably detect prohibited transformations, and how to handle ambiguous requests that could be artistic, medical, educational or abusive depending on context.

 

Minnesota Case Could Shape State AI Regulation

The ruling arrives as U.S. states experiment with increasingly specific AI laws while Congress and federal courts continue to work through broader questions about platform responsibility, synthetic media and free expression. Minnesota's approach is unusually direct because it regulates access to a category of AI capability rather than only the downstream publication of harmful content.

 

A final ruling for Minnesota could encourage other states to write provider-focused restrictions for high-risk generative features. A ruling for xAI could push lawmakers toward narrower statutes built around nonconsent, distribution, knowledge requirements or explicit safe harbors. Either outcome would affect how AI companies design nationwide products that face different state rules.

 

For now, the most important distinction is simple: xAI lost its request to pause the law, not the entire lawsuit. Minnesota can continue enforcing section 325E.91 while Frank considers the constitutional challenge, and the state's pending motion to dismiss could become the next major procedural test.

 

More on This Story