The courtroom was quiet, then the PDF hit the judge’s desk and the air changed. Two pages from the U.S. Department of Justice landed in a fight that feels bigger than nudity models or app terms. You could feel the stakes: lawmaking versus an industry that moves at the speed of code.
I’ve read the DOJ brief; you should too. This isn’t just a tech quarrel. It’s a federal voice saying state rules risk choking innovation—and public safety—with a single, sweeping sentence.
A clerk handed across a single PDF — the DOJ’s statement of interest, and it painted the law as broader than federal guardrails.
The department told the court that Minnesota’s ban on so-called nudification technology goes further than existing federal law and could create a patchwork of conflicting rules that slows industry development. It stopped short of asking the judge to grant the emergency injunction xAI sought, but its presence in the case raises the profile and legal gravity of what started as a private-company suit.
What the DOJ argued: federal statutes such as the Take It Down Act already target core harms—requiring platforms to remove reported nonconsensual intimate images, including AI deepfakes, within 48 hours after a valid request. The department’s message is twofold: federal tools exist and scattered state laws can interfere with national priorities, including economic and security competition with global rivals.
Can the DOJ influence state AI laws?
Short answer: the DOJ cannot rewrite state law, but its statement carries weight in federal court. When the department files a statement of interest, it’s a sign the federal government sees broader implications—legal precedent, interstate commerce, and national security. I read that as a warning shot rather than a final order.
On Aug. 1, Minnesota’s ban on nudification technology took effect — the statute targets platforms, not individual users.
The law bars owners and operators of websites, apps, and software services from allowing users to “nudify” images or video, and imposes civil fines up to $500,000 (€470,000) for each unlawful access, download, or use. That figure alone turned heads in Silicon Valley and state capitals.
xAI’s suit hinges on breadth: the company says the statute does not distinguish between bad actors and someone who consents, and it does not account for intent, artistic work, or parody. You can see the tension: a blunt instrument designed to prevent harm may also sweep in speech that’s legal and consensual.
What does Minnesota’s ban prohibit?
In practice, the law forbids platforms from enabling or allowing users to produce nudified images or videos. It focuses on platform liability over individual user conduct and includes steep civil penalties for violations, which in turn pushes platforms to preemptively restrict features and user content.
Outside the courtroom, xAI had already sued Attorney General Keith Ellison — the company cites First Amendment concerns.
xAI argues Minnesota’s ban is unconstitutional because it criminalizes a wide range of expression and fails to carve out consent or intent. The company points to internal policies that already forbid generating nude or sexualized images of people without consent, including minors.
Robert Dunn, xAI’s lawyer, told the court that the law could punish a platform for allowing a user to nudify themselves. Janine Kimble, defending Minnesota, replied that terms of service are not enough when users can still generate the images—a practical observation on enforcement versus promise. The state also relied on research suggesting roughly 95% of deepfakes and manipulated images are nonconsensual and sexually explicit, which framed the law as a public-protection measure.
Will this affect AI development?
Yes—if states multiply overlapping bans, companies will either stop offering features or leave states. The DOJ framed that risk as a national concern, arguing excessive state restrictions could hamper American competitiveness in AI against international rivals. For operators of platforms like Grok, ChatGPT, or Meta’s image tools, the calculus shifts from product design to legal risk — which slows product cycles and user choice.
Judge Donovan Frank refused a short-lived restraining order — the request arrived days before the law took effect.
The judge denied xAI’s request for a temporary restraining order, citing delay: the emergency plea came just days before the statute’s effective date, undermining the claim that immediate relief was necessary. He then took the request for a preliminary injunction under advisement and said he would rule as soon as possible.
The procedural posture matters. I see the DOJ’s filing as a strategic nudge that could influence how the court frames federal interest versus state prerogative. You should note: the department’s brief did not demand the injunction xAI wanted. It chose persuasion over orders, a calculated move in federal-state dynamics.
The DOJ’s filing acted like a handbrake on a speeding car—slowing down a rush into divergent state rules while lawyers argue the destination. Minnesota’s statute, on the other hand, is a neighborhood of stop signs that could make jurisdictions incompatible overnight.
If you follow platforms and policy, names matter: Elon Musk and his xAI team, Grok as the product at issue, Attorney General Keith Ellison as the state actor, Judge Donovan Frank as the arbiter, and federal statutes like the Take It Down Act as the backdrop. OpenAI and Meta watch closely; developers and compliance teams in Silicon Valley calculate new risk matrices.
I’ll be watching the judge’s ruling and the next federal moves. For you—whether you build, regulate, or use these systems—the central question is simple: who will write the rules that shape how intimate images are created, shared, and removed in the age of generative AI?