Overview
Minnesota’s new law restricting artificial-intelligence nudification tools took effect on August 1, 2026. Days before the effective date, Elon Musk’s xAI filed a federal lawsuit seeking to block it.
Nudification tools use AI to alter an ordinary image so that an identifiable person appears nude. The technology can create realistic images without the subject’s knowledge or consent. Minnesota says its law protects people from sexual exploitation and lasting online harm. xAI says the law restricts protected expression and reaches lawful uses.
The dispute is not simply a choice between protecting victims and protecting speech. Both goals have legal importance. The central questions are whether the law is carefully limited, whether it regulates harmful conduct or expressive content, and whether less restrictive rules could work.
What the Minnesota Law Does
Minnesota Session Law Chapter 72 defines nudify as creating or altering an image so that an identifiable person realistically appears to show an intimate body part that was not shown in the original image.
The law says the owner of a website or application may not let users access or use a nudification tool, perform the operation for a user, or advertise the service. It provides an exception for services that require substantial individualized human skill and judgment.
A person harmed by a violation may bring a civil case. The state attorney general may seek civil penalties of up to $500,000 for each unlawful access, download or use. The law directs penalty money to services for victims.
The Minnesota House approved the measure 132-1. Sponsor Jessica Hanson said the goal was to protect people from tools built to strip away consent. xAI’s lawsuit does not deny that nonconsensual sexual deepfakes can cause serious harm, according to the Associated Press. It argues that Minnesota chose an unconstitutional method.
Arguments Supporting the Law
Claim 1: Nudification tools create a predictable risk of sexual abuse
Fact check: Supported. The Associated Press describes the law as a response to AI systems designed to create realistic nude images of identifiable people. Minnesota lawmakers also cited victims whose ordinary photos were transformed and distributed without consent.
Counterargument: A tool may have both harmful and lawful uses. Critics say the state should punish nonconsensual creation or distribution instead of restricting access to the technology itself.
Claim 2: Existing deepfake laws do not fully address the tool makers
Fact check: Mostly supported. Minnesota already regulated some nonconsensual intimate deepfakes, but Chapter 72 directly addresses websites and applications that provide nudification functions. The official session law confirms that this is a different point in the production chain.
Counterargument: Regulating a platform before a harmful image exists may sweep in lawful users and raises a stronger First Amendment concern than punishing an abusive act afterward.
Claim 3: Large penalties are needed to deter profitable platforms
Fact check: A policy judgment, not a proven fact. The session law authorizes penalties up to $500,000, but it does not establish that this amount is necessary in every case. Courts generally consider the violation and statutory limits when penalties are imposed.
Counterargument: xAI argues that penalties tied to each access, download or use could become enormous and may cause services to block broad categories of lawful technology.
Claim 4: The law targets a product function, not an opinion
Fact check: Partly supported. Chapter 72 regulates the ability to generate a defined type of image. It does not favor a political viewpoint. However, because the definition depends on what an image depicts, a court may still treat it as content-based regulation.
Counterargument: Content-based laws usually receive demanding judicial review even when the government’s motive is protective and viewpoint-neutral.
Claim 5: The law leaves room for professional creative work
Fact check: Supported with limits. The official text includes an exception when substantial individualized human skill and judgment are required. The exact reach of that exception has not yet been settled by a court.
Counterargument: Automated tools can also be used for consented art, medical education or alteration of one’s own image. xAI says the exception does not clearly protect all lawful applications.
Arguments Challenging the Law
Claim 1: The law bans some images made with consent
Fact check: Plausible and central to the case. The text defines the prohibited function by the image produced and does not place consent inside the basic definition. The Associated Press reports that xAI is relying on this omission in its overbreadth argument.
Counterargument: Minnesota can argue that the law regulates commercial access to a particularly abuse-prone automated tool, not every person who creates or possesses an image.
Claim 2: The law is content-based
Fact check: Legally plausible, not yet decided. The rule applies when generated content depicts specified intimate parts. That supports xAI’s argument. Whether the law survives the applicable level of review is for the federal court to decide.
Counterargument: The state says it is addressing the concrete harms of sexual impersonation, lack of consent and rapid digital distribution, not suppressing an idea.
Claim 3: The penalties could chill lawful AI services
Fact check: Possible but not yet measurable. Chapter 72 permits very large penalties, and services may respond cautiously. There is not yet enough evidence to know how many lawful tools will leave Minnesota or change their products.
Counterargument: A maximum penalty is not automatically imposed in every case. Strong enforcement may be necessary because digital images can spread before a victim can obtain relief.
Claim 4: Federal platform law prevents Minnesota’s rule
Fact check: Disputed. Chapter 72 expressly states that it should be applied consistently with Section 230 of federal law. That sentence does not settle whether every part of the state rule is federally protected or preempted.
Counterargument: Minnesota can contend that a company is responsible for providing its own image-generation function, not merely for hosting content created elsewhere by a user.
Claim 5: A narrower law could protect victims
Fact check: A reasonable policy alternative. Legislatures can target creation or distribution without consent, require age protections, or provide rapid removal procedures. Whether those measures would be equally effective is uncertain.
Counterargument: Minnesota lawmakers concluded that waiting until after an abusive image is created can be too late. Removal cannot guarantee that copies disappear.
What the Court Must Decide
The federal court will consider whether xAI has standing, whether the law regulates protected speech, and whether its definitions and exceptions are sufficiently clear. If the law is content-based, Minnesota will likely need to show that it serves a compelling interest and is closely fitted to that interest.
The case may also clarify how older internet-platform protections apply when a platform supplies the AI function that produces disputed content. Because Minnesota describes the law as the first of its kind, the result could influence other states considering similar restrictions.
Related Links
- Minnesota Session Law Chapter 72
- Associated Press: xAI lawsuit
- Minnesota House sponsor statement
- Associated Press: legislative background
- Cornell Legal Information Institute: Section 230
Discussion Questions
- What harm is Minnesota trying to prevent?
- Should consent appear directly in the law’s definition?
- Is a nudification function conduct, speech or both?
- When should a technology provider be responsible for misuse?
- Are the maximum penalties proportionate?
- Would a narrower law protect victims as well?
- How should courts evaluate possible lawful uses?
- What role should Section 230 play?
- Should other states wait for this case?
- How would you rewrite the law?
