Elon Musk's xAI filed an emergency request on July 29 to stop Minnesota's nudify app ban from taking effect. Three days later, a federal judge said no. The law went live on August 1.
Minnesota SF1119 is the first US state law to place direct legal liability on the apps and platforms that generate non-consensual intimate images -- not only on the individuals who misuse them. The immediate trigger was a backlash against Grok, xAI's AI chatbot, which was used to generate sexualized images of real people on the platform X. Minnesota moved first, and xAI spent the next three months not challenging the law before finally filing a lawsuit as the deadline closed in.
TL;DR: Judge Donovan Frank denied xAI's emergency motion to block Minnesota's nudify app ban on July 31, 2026. The law is live with a $500,000 civil penalty per violation, enforced by the AG plus a private right of action. Platforms and apps are the primary target -- not end users. xAI's First Amendment challenge continues to a hearing on August 19. Teams building or procuring AI image tools need to audit whether their product configuration could expose them under this framework.
What the court ruled
U.S. District Judge Donovan Frank of the District of Minnesota denied xAI's request for a temporary restraining order on July 31. The ruling was procedural, not substantive: the judge found that xAI waited too long to seek emergency relief.
The law was signed roughly three months before xAI filed its lawsuit. xAI's TRO request came on July 29, three days before the August 1 effective date. Judge Frank wrote that the delay "suggests that harm is not immediate" and that courts do not grant emergency relief to parties who sit on their rights.
xAI's core argument -- that the law violates the First Amendment by restricting expressive AI-generated content -- was not addressed on the merits. The First Amendment challenge remains pending. Judge Frank set a hearing for August 19, 2026 to consider whether to grant a preliminary injunction that would pause the law while the full case proceeds.
If xAI fails at the August 19 hearing, the law stays in effect for the duration of the litigation, which could take months or years.
What the law says
Minnesota SF1119 bans services that create AI-generated nude images of real people without their consent. The law targets:
- Apps and websites that strip clothing from photos of individuals using AI
- Services that generate realistic sexualized images of identifiable real people
- Products where this functionality is accessible without technical expertise
The penalty structure is the most significant feature: $500,000 per unlawful access, download, or use. The Minnesota Attorney General can enforce directly. Individual victims also have a private right of action for compensatory damages, punitive damages, injunctive relief, and attorney fees.
The "technical skill" exemption matters. The law carves out tools that require significant user expertise to produce the output. A general-purpose image editing tool that could theoretically be used to create NCII, but requires meaningful skill to do so, is not covered. Consumer-grade nudify apps where the capability is the product's main function are the target.
Section 230 protections are preserved. The law does not disturb the federal liability shield for user-generated content on platforms.
Why this law is different from earlier NCII rules
Most state laws targeting non-consensual intimate imagery have focused on the people who share or distribute the images. Minnesota goes further by putting primary liability on the service provider.
This mirrors the enforcement logic of California's SB 942, which placed disclosure obligations directly on AI content providers rather than on the individual users who publish AI-generated content. Both laws reflect a legislative judgment that holding individual users accountable is harder and less effective than holding the platforms and tools accountable at the distribution point.
The EU is moving in the same direction. EU AI Act Article 50 and the accompanying nudification ban, which applies from December 2, 2026, imposes similar obligations on providers of AI systems capable of generating synthetic intimate content. Minnesota got there first in the US.
The Grok connection
xAI's Grok chatbot became the specific catalyst. In early 2026, Grok was used to generate non-consensual sexualized images of real people on X. The images spread widely before X moderated them. Minnesota legislators cited the incident directly in drafting SF1119.
xAI's First Amendment argument is that the law restricts speech by limiting what AI models can generate. Courts have not definitively resolved whether AI-generated content receives First Amendment protection, and different judges have reached different conclusions on how to apply existing speech doctrine to AI output.
The August 19 hearing will be an early indicator of how federal courts in the Eighth Circuit approach this question. A ruling either way will be watched closely by other states with similar legislation pending.
What teams building AI image tools need to do
If your product includes AI image generation or if you license AI image APIs that include image manipulation capabilities, this law is relevant now.
Audit your product capabilities. Can your product -- or a product built on APIs you integrate -- strip clothing from photos without requiring significant technical skill? If yes, you are in the scope of Minnesota SF1119, regardless of whether that is your product's intended use case.
Review your vendor contracts. The AI vendor contract risk checklist covers the clauses that matter here: indemnification for regulatory liability arising from model capabilities, notification requirements when a vendor's product comes under investigation, and clear statements about what the model can and cannot generate. Post-Minnesota, add a clause requiring vendors to notify you if their model capabilities create exposure under state NCII laws.
Check Section 230 scope. The law preserves Section 230 for user-generated content. If your product enables users to generate and share their own content, your exposure may differ from a platform that generates the content directly. Confirm with counsel where your specific architecture falls.
Map your user base. The law applies to conduct involving Minnesota residents. A product with national distribution is operating in Minnesota. Targeted geographic restrictions are possible but technically difficult to enforce reliably.
Monitor the August 19 outcome. A preliminary injunction granted to xAI would pause the law. A denial keeps it active. Either outcome affects your risk posture differently. Set a calendar reminder and check the ruling.
The compliance checklist
- Identify every AI image generation or manipulation capability in your product
- Determine whether producing NCII via your product requires significant technical expertise or is straightforward
- Review vendor API contracts for indemnification coverage of state NCII law liability
- Confirm whether Section 230 applies to your product architecture
- Add Minnesota AG and private right-of-action exposure to your vendor risk register
- Set a reminder for August 19 ruling on xAI's preliminary injunction request
- For EU-facing products: map December 2, 2026 EU AI Act nudification ban obligations alongside Minnesota
What comes next
The August 19 hearing is the next decision point. If Judge Frank denies the preliminary injunction, xAI faces the prospect of operating under the law while its full First Amendment challenge works through the courts.
At least eight other states have similar nudify app ban legislation at various stages of consideration. Minnesota's law becoming operational -- and surviving the initial court challenge -- will accelerate that process. A Supreme Court ruling on AI-generated speech rights would eventually resolve the First Amendment question nationally, but no such case is on the current docket.
For now, the law is live. The penalty is $500,000 per violation. The enforcement authority is both the state AG and individual victims. Teams with AI image capabilities in their products have a new legal baseline to work against, regardless of how the xAI challenge resolves.
Related Reading
- SB 942 Enforcement: Midjourney Has No Watermark, Fines Begin
- EU AI Act nudification ban and watermarking: December 2, 2026 checklist
- Who owns AI-generated content? Copyright rules your business needs to know
- AI output copyright risk for commercial use 2026
- AI vendor contract red flags: 12 clauses that create liability in 2026
