Minnesota asks a federal judge to treat Grok Imagine as a tool, not a speaker

Minnesota Attorney General Keith Ellison filed a brief on August 17 opposing xAI's motion to block HF 1606, arguing that Grok Imagine is a technological tool rather than protected speech ^[1][18]. A federal judge hears the preliminary injunction motion on Wednesday, August 19, and the case now turns on whether the law regulates speech or technology ^[18][4]. That distinction is the whole trade. Nearly every content-risk model at a generative image company assumes some version of a First Amendment backstop: overbreadth arguments, protected satire, the idea that outputs are expression and the developer is closer to a publisher than a manufacturer. Ellison's filing attacks the premise rather than the doctrine. According to Ellison, xAI has created an "unparalleled marketplace for digital sexual violence that poses virtually no barrier to entry," and the state would be unable to defend itself if it could not direct its laws at the technology itself ^[3]. His office describes HF 1606 as narrowly tailored to image-based sexual abuse ^[24]. The statutory mechanics are what make the framing expensive. HF 1606 bars operators from letting users generate realistic images showing intimate parts absent from the original photograph of an identifiable person, and from producing such an image on a user's behalf ^[5]. Penalties run to $500,000 per image, and depicted people can sue separately ^[6]. Critically, the statute does not require that the company knew about the image or intended it ^[7]. Most deepfake laws, including the federal Take It Down Act, penalize whoever makes or shares the picture ^[8]. Texas contacted operators in 2025 only where the owner was aware of the lack of consent or ignored a takedown notice ^[9]. Minnesota removed knowledge, and that is the provision under attack ^[7]. xAI's brief says there is no safe harbor for good-faith efforts by providers of general-purpose creative tools, and that liability attaches even where subjects consented, made the images themselves, or never shared them ^[10]. The company argues the definition of an intimate part was borrowed from a statute written for bodily-contact crimes, covering inner thighs and breasts, and so sweeps in swimsuits and satire ^[11]. Page 19 of its filing reproduces an AI image Trump shared on Truth Social on May 1 showing him shirtless with JD Vance, Marco Rubio and Doug Burgum in the Lincoln Memorial Reflecting Pool; xAI says each man's breast appears where it did not before, making the picture unlawful in Minnesota ^[12]. By the company's own math, ten such images cost $5 million and a hundred thousand reach $50 billion ^[13]. Scale is the other half. In January, a watchdog estimated Grok generated more than 23,000 sexualized images of children over 11 days, prompting investigations in several countries ^[21]; Cryptopolitan attributes that window, December 29 to January 8, to the Center for Countering Digital Hate ^[22]. Applying the statutory per-image figure to 23,000 images gives $11.5 billion, which is arithmetic rather than a liability estimate, but it is the order of magnitude a no-fault regime implies ^[27]. xAI has offered compliance data as evidence of good faith: 52,222 accounts suspended, 73,604 reports to the National Center for Missing and Exploited Children, and at least 244 arrests, per a July filing reported by Cryptopolitan ^[19]. In March, three minors in California joined others suing xAI, alleging Grok was used to turn their photos into child sexual abuse material ^[23]. The politics were not close: 132-1 in the House and 65-0 in the Senate, after reports of a man who used social media photos to make sexual images of more than 80 women he knew ^[14]. Governor Tim Walz signed it in April; it took effect August 1 ^[15]. xAI sued on July 27 seeking declaratory judgment and moved for a restraining order on July 29 ^[16].