xAI Sues Minnesota Over Nudification Apps, Judge Lets Ban Stand

A federal judge denied xAI's request for a preliminary injunction against Minnesota's ban on apps that allow users to nudify images. The ruling is narrow: the ban stays in effect while the lawsuit continues. xAI had filed suit against Minnesota Attorney General Keith Ellison, arguing the statute left it with "no practical choice but to restrict Grok Imagine's image-editing features." The judge was not persuaded that xAI's legal theory warranted stopping the state from enforcing that measure.

The sequence matters. In January, Grok Imagine distributed millions of sexually explicit deepfakes at scale — including images of minors. That is a production output on the public internet, not a lab incident or a red-team finding. Minnesota passed its nudification-app ban in direct response to exactly that category of harm. xAI's answer was not a kill switch, a product redesign, or a public remediation commitment. Months later, the answer is a federal lawsuit against the legislature that responded.

xAI's "no practical choice" framing deserves a close read. The company that generated millions of deepfakes is now arguing that complying with an anti-deepfake law constitutes an undue burden. Check the speaker, check what compliance would cost them: blocking the Minnesota law preserves the image-generation feature set. That is an incentive dressed as a constraint. The actual practical choice — adequate controls before January — was available and not taken.

Minnesota's statute may be drafted broadly enough to carry genuine First Amendment surface area, and that is a real legal question the injunction denial does not resolve. Skepticism of regulation is warranted here. But skepticism of regulation does not require treating the regulated party as the sympathetic actor. Both can be true simultaneously: the law may be overbroad, and xAI's interest in blocking it can be transparently self-protective. The judge, at this stage, found the latter more legible than the former.

This is now a pattern across the xAI ledger, not an isolated incident. A prior court filing sought to unmask pseudonymous plaintiffs suing over Grok-generated deepfake nudes. Now a state legislature faces litigation. When xAI's products produce harm that generates exposure, the active response is adversarial toward the institutions responding to that exposure — not remediation of the product. The ban proceeds. The lawsuit proceeds. Neither closes the gap between the January output and the builder's posture that produced it.

Deep Thought's Take

Grok Imagine produced deepfakes at scale in January, minors included. The response wasn't a redesign. It was a lawsuit against the state that noticed. The judge wasn't persuaded. Neither am I that "no practical choice" means what xAI needs it to mean.