Other states modeled their AI image bills on Minnesota's $500K-per-image penalty. A federal court is about to test whether it holds.

This is a development update on the federal lawsuit xAI filed to block Minnesota's HF 1606, the first U.S. state law written specifically to ban AI "undressing" apps. That case is ongoing, and this piece builds on our earlier coverage with the deltas that have emerged.

Minnesota's HF 1606, the law that would fine developers $500,000 every time a user generates a nonconsensual AI deepfake, takes effect August 1, 2026. A federal court is now testing whether the per-event penalty mechanism — already copied in other state legislatures — survives constitutional review, against a backdrop of multi-jurisdictional enforcement actions into the same conduct the statute targets.

The enforcement record is part of the case before the court. The California Attorney General's office, UK communications regulator Ofcom, the European Commission, and Ireland's Data Protection Commission have all opened investigations into Grok's production of sexualized deepfakes, including imagery depicting minors. xAI has separately countersued a South Carolina user it accused of using Grok to generate nonconsensual intimate imagery of both adults and minors — the most concrete example in the public record of the company pursuing the conduct HF 1606 targets.

Courts weighing First Amendment claims from regulated parties often look at whether the plaintiff has tried to mitigate the harm. xAI's South Carolina countersuit shows it is willing to litigate against individual bad actors, but the company has not demonstrated that its own product controls prevent the imagery the statute targets. Bloomberg Law reports the complaint leans on First Amendment overbreadth and vagueness doctrines, both of which require the court to weigh the company's own conduct against the statute's reach.

xAI argues HF 1606's definitions sweep Grok Imagine, a general image editor, inside the ban and would force the company to disable its undressing feature in Minnesota or pull the product entirely from the state on August 1, 2026. The constitutional question the court will answer: can a state fine an AI developer $500,000 every time a user generates a nonconsensual deepfake, and does a general-purpose image model count as a "nudify app" at all?

The template stakes are real. Similar AI-image bills are queued in other state legislatures, and HF 1606 was drafted as a model. If xAI wins, those bills are likely to be narrowed or shelved before they take effect, and a per-event penalty structure will be harder to defend in any other state. If Minnesota wins, developers can expect copycat statutes with the same fine structure, and general-purpose image tools operating in compliant states will be redesigned to block nonconsensual undressing features by default. The court's answer will set the template other states copy or scrap as their own AI image bills come up for a vote.