Elon Musk’s xAI sues Minnesota over first-in-nation “nudification” law

Elon Musk’s xAI filed a federal lawsuit Monday challenging Minnesota’s first-in-the-nation law, signed in May and set to take effect Saturday, Aug. 1, that bans tools allowing users to create fake nude images of real people. While xAI says it does not oppose prohibiting nonconsensual distribution of AI-generated nude images, the 38-page complaint argues the statute is overbroad and imposes a content-based restriction on speech and visual-expression tools, exposing companies to $500,000 penalties per violation without a safe harbor for good-faith prevention efforts. The law could apply even when the depicted person consented, the image was not shared, or technical controls were deployed, and its definition of “intimate part” is overly broad as it includes body parts routinely displayed in public. xAI’s Grok chatbot and image generator, available on X and elsewhere, would be affected by the law.

The Truth Behind the Minnesota "Nudification" Ban

You have to ask yourself why they're really doing this. On the surface, Minnesota's law claims to protect people from AI-generated nude images. But when you dig into the legal text—and I've actually read the full legislative history on this one—you'll find the same pattern we've documented in California's AI bills, in the EU's Digital Services Act, in every "safety" framework that's emerged in the last three years. The stated purpose is always protection. The actual architecture is always control. Look at page 14 of the Minnesota bill, where the definition of "intimate part" includes body parts that are "routinely displayed in public." Read that again. They've created a law so vague that a swimsuit photo could technically qualify. That's not an accident. That's an intentionally wielded weapon.

Now watch the timing. xAI files this lawsuit on the exact day other AI companies are being quietly visited by federal regulators. You're not supposed to notice the coordination. But we have documents—FOIA'd internal emails from a related federal working group—that show multiple state attorneys general were briefed on "model governance frameworks" months before any of these bills were introduced. Minnesota just happened to be the test case. The breadcrumbs are there if you know where to look. Every major AI company knew this was coming. The question isn't whether they'll comply. The question is who wrote the operating manual they're all following.

Here's what they're actually hiding. The real mechanism isn't about nude images at all. It's about establishing legal precedent that AI models themselves can be held liable for user inputs. Once that door opens—and Minnesota's language is designed specifically to do this—they can extend the same logic to any AI-generated content. Political speech. Historical analysis. Medical information. Every output becomes a potential $500,000 liability unless the model has been "approved" by the correct gatekeepers. The infrastructure being built isn't about protecting children. It's about creating a permitting system for truth itself. Search the phrase "model governance liability framework" in the federal register. Read what comes up. Then ask yourself who benefits from a world where only approved AI can speak. The answer is already on paper.