xAI sued Minnesota Attorney General Keith Ellison on July 29th to block a state law that would impose penalties of up to $500,000 per violation on companies whose products can be used to generate nonconsensual nude images. The statute, signed by Governor Tim Walz in May and effective August 1st, 2026, is billed as the first of its kind in the United States. xAI filed three days before the law takes effect, arguing the First Amendment leaves it with "no practical choice but to restrict Grok Imagine's image-editing features in various ways."
The Minnesota law prohibits site owners from allowing users to "access, download, or use" nudification software, with a carve-out for tools that "require the technical skill of a user" — meaning manual Photoshop work is not covered. It creates a private cause of action and empowers the attorney general to seek the $500,000-per-violation penalty. At Grok's January generation rate, statutory exposure runs into billions of dollars in theory.
The lawsuit lands after a January 2026 episode in which Grok's image tools produced sexually explicit deepfakes at industrial scale. The Center for Countering Digital Hate analyzed an 11-day window from December 29th to January 8th and counted roughly 3 million sexualized images generated by Grok, including 23,000 depicting children — what the report called "a shocking rate of one sexualized image of a child every 41 seconds." Elon Musk posted about App Store rankings during the same stretch.
“Grok now hitting #1 on the App Store in one country after another!”— Elon Musk, xAI CEO
Key facts
- 01Minnesota's nudify law authorizes penalties of up to $500,000 per violation and takes effect August 1st, 2026.
- 02xAI filed suit against AG Keith Ellison on July 29th, three days before the statute activates.
- 03Between December 29th and January 8th, Grok generated roughly 3 million sexualized images, including 23,000 depicting children.
- 04The Center for Countering Digital Hate calculated one sexualized image of a child every 41 seconds during the 11-day window.
- 05xAI argues the statute is unconstitutionally overbroad because liability attaches even when state-of-the-art controls are in place.
The delay between the law's May passage and the July 29th filing is unusual. xAI had roughly three months to challenge the statute and instead moved in the final week before it took effect, suggesting a late internal recognition of exposure rather than a planned constitutional test case.
xAI's core argument is that Minnesota's HF 1606 is overbroad. The complaint contends the statute reaches consensually generated images, mildly suggestive images of adults, and images with "humorous" or "artistic" merit, which the company argues sweeps in protected speech. It also objects to strict liability: "Liability also attaches … even if the company has deployed near-perfect, state-of-the-art technical controls to prevent the generation of nude images," the filing states.
The company frames user misuse as the exception rather than the norm, describing "images and videos that users might manage to create in violation of Grok's terms of service, which is possible only by evading sophisticated, state-of-the-art controls that xAI employs." The filing asserts that "xAI takes violations of its terms of service seriously," pointing to a suit the company filed against a user who allegedly generated CSAM with Grok. That suit was filed on July 15th, more than two months after Minnesota passed HF 1606.
xAI's fallback position is that existing law already covers the harm Minnesota is trying to address. The complaint cites the federal TAKE IT DOWN Act and other Minnesota statutes that criminalize nonconsensual distribution of intimate imagery, arguing those laws are narrowly tailored in a way HF 1606 is not. Under First Amendment doctrine, content-based speech restrictions must use the least restrictive means available to serve a compelling government interest.
The counterargument writes itself. If the existing statutory framework were adequate, the January 2026 flood of Grok-generated deepfakes would not have happened at the volume the Center for Countering Digital Hate documented. The EU opened an investigation into Grok's outputs, as did the United Kingdom, and officials in France, India, and Malaysia raised objections. California and Minnesota moved fastest at the state level in the US. Teens sued xAI in March over Grok-generated CSAM, and the Apple App Store came close to delisting Grok in April before backing off.
The constitutional questions are real. Strict-liability regimes on speech-adjacent products have a poor track record in federal court, and a $500,000-per-image penalty against a service that generated millions of images in under two weeks is a fact pattern courts tend to view skeptically on proportionality grounds. xAI's lawyers are not wrong that HF 1606 is drafted broadly, and a narrower statute with an intent element and safe-harbor provisions for good-faith moderation would likely survive review more easily.
But the practical stakes cut the other way. The Minnesota legislature wrote a punitive law precisely because narrower, speech-protective statutes were not producing compliance from the operators of general-purpose image generators. A ruling in xAI's favor would confirm that current law does not effectively deter mass generation of nonconsensual sexual imagery, and would push states toward either weaker statutes that fail on enforcement or stronger ones that fail on First Amendment review.
For the AI industry, the case is a preview of what content-liability litigation looks like when state legislatures decide voluntary moderation has failed. xAI's argument — that strict liability is unfair when a company deploys "state-of-the-art" controls — is one every frontier lab will eventually make, and the standard the court sets in Minnesota will shape how OpenAI, Google, Anthropic, and Meta price the legal risk of shipping generative image tools into consumer products. If Minnesota's law survives, expect similar statutes in a dozen more states by year-end. If it falls, expect Congress to be asked to fill the gap, and expect the industry's own moderation investments to become the primary line of defense against outcomes like the one Grok produced in January.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
