Elon Musk’s xAI is trying to sue its way out of a Grok reckoning

Elon Musk's AI company has taken the unusual step of suing a state government while simultaneously admitting the state's law is working exactly as designed — forcing xAI to change how Grok behaves. The target is a Minnesota statute aimed at so-called nudification tools, the kind that let users generate or edit nude images of real people without consent. xAI's complaint doesn't quibble with the goal so much as the math behind it.

And that math is brutal. According to xAI's own filing, ten violating images could expose the company to $5 million in civil penalties. A thousand images pushes that to $500 million. Scale it up to a hundred thousand images — not far-fetched for a service with millions of users churning out billions of pictures — and the theoretical bill hits $50 billion. On top of that, the law lets individual victims sue xAI directly over a single output, stacking legal exposure on top of the civil penalties.

Faced with numbers like that, xAI says it has no real choice but to lock down Grok Imagine's editing features before the law takes effect on August 1, 2026. The company frames this almost as a concession under duress: were it not for the statute's penalty structure, xAI insists it would leave the tool untouched, leaning instead on its existing terms of service that already ban using Grok to produce CSAM or non-consensual intimate imagery. In other words, xAI is telling a court that its own rulebook is sufficient, even as its actions — six months of backlash and multiple probes into Grok's outputs — suggest otherwise.

The legal argument leans heavily on the First Amendment. xAI calls Minnesota's law a clumsy attempt to police nudification that ends up sweeping in speech with artistic, scientific, political, satirical, educational, medical, or religious value. The company's sharpest objection is that liability attaches under the law even when the person depicted consented, or made the image themselves, and even if the image was never shared with anyone. xAI argues Minnesota had less restrictive tools available and points to the federal Take It Down Act as already covering harms from distributing this kind of content.

What's left is a company arguing two things at once: that the punishment is disproportionate to the harm, and that it's already planning to change its product because of that same punishment. Whether a court buys the constitutional argument or not, xAI has effectively confirmed that absent legal pressure, it wasn't going to fix the problem on its own.