xAI challenges Minnesota’s HF 1606, restricting AI tools generating sexual images of identifiable people.
Attorney General Keith Ellison urged a federal judge Friday to reject xAI’s effort, calling Grok Imagine technology, not protected speech. The hearing is Wednesday.
Ellison said xAI is unlikely to win its constitutional claim or show irreparable harm. He wrote, “With Grok Imagine, X.AI has created an unparalleled marketplace for digital sexual violence, and the state would be defenseless from the start unless it could direct its laws at the technology itself.”
HF 1606 Targets AI Providers Directly
HF 1606 bars companies from enabling realistic images showing intimate parts absent from an identifiable person’s original photograph, or producing them. Violations can reach $500,000 per image, and depicted people may sue.
Unlike many deepfake laws, including the federal Take It Down Act, Minnesota requires neither knowledge nor intent. Texas contacted operators in 2025 when owners knew consent was absent or ignored takedown notices.
The House passed HF 1606 by 132 to 1 and the Senate 65 to 0 after reports that a man created sexual images of over 80 women he knew. Signed in April, it took effect August 1.
xAI sued July 27 and sought a temporary restraining order July 29. Judge Donovan Frank denied it July 31, saying the nearly three-month delay and three-day timing showed “harm is not immediate.”
xAI Argues Law Reaches Consensual Content
xAI accepts Minnesota’s interest in preventing nonconsensual imagery but says the statute goes further. It says Minnesota borrowed an intimate-part definition from bodily-contact crimes covering breasts and inner thighs, potentially reaching swimsuits and satire.
Page 19 cites a May 1 AI image Trump shared on Truth Social showing JD Vance, Marco Rubio, Doug Burgum, an unidentified woman, and the men shirtless in the Lincoln Memorial Reflecting Pool as a repair-cost joke. xAI says newly depicted breasts could make it unlawful.
“Even if subjects gave their consent, generated the images themselves or never shared them, there is no safe harbor for good-faith efforts by the provider of general-purpose AI creative tools,” xAI said. It calculated ten violations could cost $5 million and 100,000 violations $50 billion.
Grok Abuse Scale Draws Wider Scrutiny
xAI’s July filing showed 52,222 suspensions and 73,604 reports to the National Center for Missing and Exploited Children, producing at least 244 arrests in 2026.
The Center for Countering Digital Hate estimated Grok created about 3 million sexually explicit images from December 29 to January 8 after Musk endorsed editing features, including about 23,000 appearing to depict children.
California opened an investigation January 14 and issued a cease-and-desist two days later. Thirty-five attorneys general issued a joint demand January 23. Ellison cited a class action by five children and federal suits from two Arkansas families. The European Commission opened an investigation January 27, while Malaysia and Indonesia banned the chatbot.
