Minnesota’s Nudification Ban Capable of Punishing People

The fight over Minnesota’s “nudification” ban is not just about stopping AI‑generated sexual abuse; it is an early test of whether states can regulate powerful image tools without criminalizing parody, art, and even stripped‑down selfies in the process.

Key Points

Minnesota’s H.F. 1606 makes it illegal for websites, apps, and software to let users “nudify” images or videos of identifiable people, with civil penalties up to $500,000 per use.
The law’s definition of “nudification” reaches any realistic AI or software‑generated depiction of an intimate body part not present in the original image, including inner thighs and shirtless torsos.
Elon Musk’s xAI has sued, arguing the statute is an overbroad, content‑based restriction on free speech that covers consensual and satirical imagery alongside harmful deepfake porn.
Supporters counter that the law is narrowly aimed at nonconsensual sexualized impersonation and regulates platform conduct, not art or general image editing.
The case sits inside a broader global struggle to curb nonconsensual intimate image abuse while preserving First Amendment protections and innovation in generative AI.

What Minnesota’s Nudification Ban Actually Does

Minnesota’s law is built around a simple but far‑reaching premise: if your service lets ordinary users turn someone’s photo into a realistic nude or sexualized image, you cannot offer that capability to people in the state. The statute, codified as H.F. 1606 and added to Minnesota’s obscenity chapter, prohibits any person who “owns or controls a website, application, software, program or other service” from allowing a user to access, download or use it to “nudify” an image or video, or to perform nudification on a user’s behalf. Advertising or promoting nudification services is banned as well.

Legally, “nudify” means altering or generating an image or video to depict an intimate part of an identifiable individual that was not shown in the original, where the result is realistic enough that a reasonable viewer would believe the intimate part belongs to that person. “Intimate parts” are defined broadly: primary genital area, groin, inner thigh, buttocks, or breast. Minnesota did not limit this to fully nude pornography; it explicitly includes partial nudity and suggestive framing.

The enforcement scheme is civil rather than criminal, but it is aggressive. The Attorney General can seek a civil penalty of up to $500,000 per unlawful access, download, or use, and victims depicted in nudified images can sue for compensatory damages (including mental anguish), punitive damages, injunctive relief, attorneys’ fees, and other equitable remedies. For a platform with millions of image operations, these numbers are potentially ruinous.

Why Lawmakers Went This Far

The law did not emerge from nowhere. It was prompted by concrete cases in which individuals used apps to strip clothing from images of women, including one Minnesota incident involving more than 80 victims whose social media photos were turned into sexualized images without consent. Victims described severe psychological and social fallout, with one woman saying it affected nearly every aspect of her life. Legislators treated this as a distinct harm category, not just an extension of traditional privacy law.

HF 1606 passed with overwhelming bipartisan support—reported votes of 132–1 in the House and 65–0 in the Senate—reflecting a strong judgment that the problem was serious and urgent. Sponsors emphasized that existing state and federal rules against “revenge porn” and nonconsensual explicit deepfakes were not keeping pace with the ease and scale of AI nudification tools. They framed the bill as closing a specific loophole: instead of punishing individual uploaders after the fact, the law directly targets the platform features that make mass nudification possible.

Supporters also pointed to research on nonconsensual intimate image abuse showing how quickly synthetic nudes can spread and how difficult removal is once they are embedded in social platforms and search indexes. This aligns with a broader policy trend away from purely victim‑initiated takedown remedies and toward imposing duties on intermediaries—platforms, app stores, and tool providers—to prevent certain categories of harm at the source.

xAI’s Lawsuit: Overbreadth and Free Speech

Into this environment stepped xAI, Elon Musk’s artificial intelligence company, which operates the Grok chatbot and image generator. Days before the law took effect, xAI filed a 39‑page complaint in federal court against Minnesota Attorney General Keith Ellison, arguing that HF 1606 is unconstitutional. The complaint describes the statute as a “content-based ban on free speech and the tools of visual expression,” asserting that it imposes an overbroad restriction on protected images as well as harmful ones.

Crucially, xAI does not deny Minnesota’s interest in stopping nonconsensual nude deepfakes. The company explicitly concedes that banning distribution of AI-generated nude images of real people without consent is a legitimate state objective. Its argument is about scope: that the law “extends far beyond that goal” and covers images a person consents to, images a person creates of themselves, and satirical or artistic depictions of public figures.

As examples, commentators sympathetic to xAI’s position point to satirical or political imagery that could be swept in. Technology writers and outlets such as Reason and Techdirt have noted that the statutory language is broad enough to reach, for instance, a South Park–style parody video of President Trump wandering nude through a desert, or a fake image Trump himself posted of him and Vice President Vance shirtless in the Lincoln Memorial reflecting pool—because these depict identifiable individuals with realistic, partially nude bodies.

xAI also highlights the statute’s impact on platforms trying to moderate responsibly. The complaint stresses that HF 1606 contains no safe harbor for “good-faith moderation efforts,” meaning that a service could still face massive penalties even if it deploys filters, human review, and reporting tools to discourage misuse. The company argues that the only way to avoid risk is to disable broad categories of image editing and generation features for Minnesota users—creating a chilling effect on legitimate expression and experimentation.

How Narrow Is the Statute, Really?

Supporters of Minnesota’s law respond that these fears are overblown, pointing to several narrowing features in the statute’s text and legislative summaries. First, the law only applies when the altered or generated image is realistic enough that a reasonable person would think the intimate part belongs to the identifiable individual. Cartoonish caricatures, crude photoshop collages, or obviously fantastical compositions are arguably outside its reach.

Second, HF 1606 includes an exemption for technology that requires “the substantial application of technological or artistic skill by a human creator directing and controlling the output.” In plain terms, if producing a nudified image demands significant manual editing or professional expertise, the law treats that differently from one‑click or prompt‑driven AI services offered to the general public. Lawmakers framed this carve‑out as distinguishing bespoke studio work from automated nudify apps.

Third, enforcement targets platform operators and software vendors, not individual users. The statute prohibits people who own or control websites or apps from allowing access to nudification tools, and allows suits against those who “promote” or “advertise” such features. This is consistent with a prevention strategy focused on the distribution infrastructure; Minnesota is not criminalizing possession of altered images as such, but making it economically and legally hazardous to run services designed for nudification.

Finally, HF 1606 creates only civil liability—no one goes to jail under this law. The penalties are substantial, yet they sit within a civil enforcement framework that courts can tailor, reduce, or enjoin if applied abusively. Proponents argue this is less draconian than criminal obscenity bans, and that judicial oversight over damages and injunctive relief provides an additional safeguard.

Where the Real Legal Tension Lies

Despite these narrowing moves, the available public record does not yet contain a detailed clause‑by‑clause defense of the statute against the specific free speech objections xAI raises. Legislators emphasize the harms and insist the law “regulates conduct, not content,” but there is not yet a judicial opinion adopting that framing or addressing how HF 1606 interacts with First Amendment doctrines on content‑based restrictions, overbreadth, and prior restraint.

This is where the case becomes important beyond Minnesota. U.S. courts have historically treated sexually explicit deepfakes and revenge porn as conduct that can be regulated without violating free speech, especially when laws focus on nonconsensual distribution and include clear intent requirements. However, they also scrutinize statutes that reach truthful or satirical depictions of public figures, consensual adult expression, and tools that can be used for both lawful and unlawful purposes.

The Minnesota law sits at that intersection. On one side are victims of nonconsensual intimate image abuse, whose experiences show that existing remedies are often slow, fragmented, and ineffective in the face of rapidly evolving AI tools. On the other side are platforms and creators who worry that expansive definitions of “nudification,” tied to realistic depictions of any intimate part, can encompass benign or socially valuable expression: documentary reconstructions, political satire, gender‑affirming art, or even consensual erotic experimentation shared among adults.

Early judicial signals are mixed. A federal judge has already declined xAI’s request for a temporary injunction to pause the law, finding limited risk of immediate irreparable harm before full merits briefing. That does not resolve the constitutional question, but it suggests the court did not view xAI’s arguments as so strong that the law needed to be frozen before it even took effect. As the case progresses, more granular analysis of HF 1606’s tailoring will emerge.

Minnesota's anti-nudification law could punish people over parodies and stripped-down selfies. Elon Musk's xAI is suing. https://t.co/zPFDkN7Czm

— reason (@reason) August 5, 2026

Broader Implications for AI, Platforms, and Speech

Whatever happens in xAI v. Minnesota, the underlying conflict is not going away. Generative AI has made it trivially easy to produce synthetic intimate images of real people—friends, classmates, politicians, celebrities—at scale, and empirical work on nonconsensual intimate image abuse shows the damage can be profound and enduring. At the same time, the same tools power creative remix, satire, and legitimate sexual expression that many adults regard as protected and important.

States and national governments are experimenting with different regulatory levers: creating civil causes of action for victims, expanding criminal statutes to cover deepfake pornography, and, increasingly, imposing duties on intermediaries to prevent certain types of content from being generated or distributed in the first place. Minnesota’s nudification ban is one of the most aggressive U.S. examples of that last approach, explicitly targeting the tools rather than only punishing users.

The stakes are high for platforms. Heavy civil penalties, combined with ambiguous definitions, can produce overblocking—companies disable or severely limit image features to avoid legal risk, thereby shrinking the expressive possibilities for users who would never create nonconsensual nudes. At the same time, weak or delayed regulation leaves victims with little practical recourse against waves of synthetic sexual content that can be generated and re‑uploaded indefinitely, often from anonymous accounts.

The long‑term policy challenge is to design statutes that: (1) specifically target nonconsensual sexualized impersonation of identifiable individuals; (2) include clear consent, intent, and realism thresholds; (3) distinguish automated nudification services from general‑purpose editing tools; and (4) provide safe harbors for platforms that adopt robust, verifiable safeguards while preserving protected expression. HF 1606 is an early, imperfect attempt to meet that challenge. xAI’s lawsuit forces courts—and, indirectly, future legislators—to confront where the line between harm prevention and censorship should be drawn.