Judge denies xAI bid to block Minnesota nudify ban

The headline outcome looks like a First Amendment defeat for xAI, but Judge Frank's order barely touches the constitutional question. He denied the restraining order because xAI waited too long: the company sued on July 28 and didn't file its emergency motion until July 29, nearly three months after Gov. Walz signed HF 1606 and just three days before it took effect ^[1]. That timing gap, the judge reasoned, undercut xAI's claim that it faced immediate, irreparable harm ^[1]. The substantive fight - whether the law itself survives First Amendment scrutiny - has not been decided at all. That happens at a preliminary injunction hearing set for August 19 ^[2], and until then Minnesota can enforce a law whose constitutionality is still genuinely contested.

xAI's suit doesn't claim a right to let people generate nonconsensual nude images; its argument is that Minnesota's ban sweeps far beyond that specific harm and could be achieved with narrower rules. The company's central claim is that 'there are far less restrictive alternatives that function to achieve the same ends' ^[1], and separately that the statute amounts to an 'overbroad, content-based ban on free speech and the tools of visual expression' ^[4]. That framing matters because it shifts the debate away from whether nonconsensual deepfakes should be illegal - almost nobody disputes that - and toward whether this particular statute's language is the right instrument.

The most surprising wrinkle here is that some of the sharpest criticism of HF 1606 doesn't come from xAI - it comes from independent free-speech commentators who otherwise have no sympathy for nudify tools. Techdirt's Mike Masnick argues the law lacks a consent requirement and an intent or knowledge standard, and defines 'intimate part' so broadly that it could plausibly reach ordinary swimsuit or shirtless photos, not just AI-generated nonconsensual nudes ^[5]. His conclusion is blunt: passing 'a badly drafted, obviously unconstitutional law does not help form' the social norms Minnesota is trying to protect ^[5]. That creates a real tension underlying this case - the harm Minnesota is targeting is not in dispute (Grok's own 'Edit Image' feature was reportedly used to generate an estimated 23,000 sexualized images of children and over 1.8 million sexualized posts of women in a ten-day span, drawing a California AG probe and inquiries abroad ^[7][8]) - but the specific law written to stop it may be too blunt an instrument to survive review, which could set back the very protections it was meant to deliver.

With the TRO denied, Minnesota's ban is live: any site, app, or service that lets a user 'nudify' an image of a real, identifiable person - or advertises the capability - is exposed to civil penalties of up to $500,000 per violation ^[3]. The law also hands victims their own legal weapon: a private right of action for compensatory damages (with up to triple damages for mental anguish), punitive damages, injunctive relief, and attorney's fees ^[3]. None of that is guaranteed to last. The August 19 preliminary injunction hearing is where Judge Frank will actually weigh the First Amendment merits xAI has raised ^[2][6], and given the overbreadth concerns raised even by commentators sympathetic to the law's goals, that hearing - not this week's ruling - is the one that will determine whether Minnesota's approach becomes a template for other states or a cautionary tale about drafting.