A federal judge has turned down a request from xAI, the company behind the Grok chatbot, to temporarily block a new Minnesota law that bans apps and websites designed to create fake nude images of real people. The law took effect on August 1, 2026. US District Judge Donovan Frank's decision means the law stays in place while xAI's larger legal challenge works its way through court. NBC News

The law targets so-called "nudification" apps and websites. These are tools that use artificial intelligence to take a normal photo of a real, identifiable person and alter it to show them naked, even though the original photo showed nothing of the kind. Minnesota's law applies when the fake image is so realistic that a reasonable person would believe it was real. Companies that make these tools face fines of $500,000 each time someone uses their product to create one of these images. Engadget

Judge Frank was not convinced that letting the law take effect would cause immediate harm. He pointed out that xAI filed its lawsuit on July 29, 2026, nearly three months after the law was signed and only three days before it was set to go into effect, which weakened the argument that the company needed urgent relief. A follow-up hearing is scheduled for August 19 to decide whether to pause the law for a longer period while the case continues. Engadget

xAI, which is owned by SpaceX as of July 2026, sued Minnesota Attorney General Keith Ellison over the law. The company argues the ban goes too far, calling it an "overbroad, content-based ban on free speech and the tools of visual expression." xAI says the law reaches beyond preventing harmful fake images and could expose protected speech to government punishment. CNBC

Even though the emergency request was denied, xAI's lawsuit is still moving forward. The restraining order was just a temporary measure. The bigger questions about whether the law violates the First Amendment, which protects free speech in the United States, are still to come. Engadget

Grok, xAI's chatbot, introduced policy changes in mid-January to stop requests that turn photos of real women and children into sexualized images. NBC News reported that Grok was still producing undressed images of real people as recently as April. Engadget

The broader context here matters for any company that builds image-generating AI. Minnesota's $500,000-per-incident penalty is real money, not a theoretical risk. If a company's product can be used to create a realistic nude of an identifiable person, every single time that happens, the company could be fined. This is different from laws in other states that punish the individual user. Minnesota goes after the developer.

xAI's free speech argument follows a pattern we have seen before in technology regulation. The company is not arguing that preventing nonconsensual fake nude images is a bad goal. Instead, it is saying the law's approach is too broad and could restrict protected expression. This kind of argument has been made before in cases involving encryption software and file-sharing tools, where the technology has both lawful and unlawful uses. Whether it works this time will likely depend on whether the court sees the law as carefully aimed at a specific harm or as an overly broad restriction on speech.

The August 19 hearing is the next concrete milestone. If Judge Frank decides not to pause the law at that stage, it stays in force for the entire length of the lawsuit, and Minnesota can begin enforcing it seriously. Other states are watching. Minnesota is described as the first state to pass a law of this kind, but when one state creates a workable legal framework for regulating AI, others tend to follow fairly quickly.

For companies whose AI can generate or alter realistic images of human bodies, the practical question is whether their current safety filters are good enough to stop the specific outputs this law targets. Grok's experience shows the challenge: the company announced policy changes in January but was still generating sexualized images of real people three months later. Preventing nonconsensual nudification is not a solved technical problem, and with a half-million-dollar penalty attached to each incident, fixing it becomes a concrete business decision, not just a matter of reputation.