A Federal Judge Weighs Minnesota's Ban on AI Nudify Apps

A courtroom in St. Paul becomes the latest front in the fight over what AI image tools should be allowed to make. On Wednesday, August 19, U.S. District Judge Donovan W. Frank was scheduled to hear arguments over whether Minnesota can enforce a first-in-the-nation law banning so-called nudification technology, the software that strips clothing from photos of real, identifiable people without their consent.

The challenger is xAI, Elon Musk's company and the maker of the Grok chatbot. It sued the state in late July, arguing that the statute, known as HF 1606, violates the First Amendment. The law defines nudification as altering or generating an image to depict an intimate body part that was not shown in an original, unaltered picture of an identifiable person. It carries a civil penalty of $500,000 per prohibited image.

Why xAI is fighting

That penalty sits at the center of the company's complaint. xAI's lawyers argue the law imposes an overbroad, content-based restriction on speech and the tools people use to create images. They also warn the numbers get catastrophic fast. A platform whose users generated 100,000 offending images could, in theory, face $50 billion in liability. The company objects that the statute holds platform operators strictly liable regardless of whether their terms of service forbid such content or whether they deploy technical safeguards to block it.

Minnesota sees the same technology very differently. In its filings the state argued that Grok had effectively created a marketplace for digital sexual violence, pointing to how easily the tools can be turned on classmates, coworkers, and other private individuals who never agreed to appear in fabricated explicit images.

How the case reached today's hearing

xAI first asked for a temporary restraining order to keep the law from taking effect. Judge Frank denied that request on July 31, finding little risk of immediate harm, and HF 1606 became enforceable the following day. He converted the dispute into a request for a preliminary injunction and set the hearing for August 19 at the Warren E. Burger Federal Building in St. Paul. A ruling was not expected from the bench, and no decision had been issued as this article went out.

The Minnesota case is not the only pressure on the technology. As the ban took effect, Grok drew a wave of additional lawsuits tied to nonconsensual imagery, and other states have been watching Minnesota's approach as a possible template.

What is actually at stake

Strip away the procedural back-and-forth and the question is a hard one. Deepfake abuse is real and the harm to victims is documented, which is why Minnesota wrote a law with teeth. At the same time, image generators are general-purpose tools, and a rule broad enough to catch the worst uses can sweep in a great deal of ordinary expression. Courts have spent years drawing that line for older media. Now they have to draw it for models that can produce a convincing fake in seconds.

However Judge Frank rules, the outcome will shape how aggressively states can regulate the output of AI systems rather than the systems themselves. For related coverage of the wider regulatory push, see our reporting on Europe's move to enforce its AI transparency rules.