xAI sues Colorado
Elon Musk’s xAI is suing the state of Colorado, arguing the new state AI law violates the First Amendment and will stifle innovation. The complaint formalizes industry pushback against state-level regulation and marks a high-profile legal test of how far states can go in policing AI. (x.com)
Colorado is about to start enforcing the first broad state law in America aimed at how companies build and use high-risk artificial intelligence, and Elon Musk’s xAI went to federal court on April 9 to try to stop it before the June 30 start date. The company says Colorado is not just regulating products but dictating what an artificial intelligence model is allowed to say and how it must be trained. (reuters.com) The Colorado law is Senate Bill 24-205, signed on May 17, 2024, and it covers “high-risk” systems used to help make life-changing decisions in jobs, housing, lending, education, insurance, health care, and legal services. It tells developers and deployers to use reasonable care to prevent algorithmic discrimination against consumers. (leg.colorado.gov) Algorithmic discrimination is the idea that a computer system can sort people unfairly the way a biased manager or loan officer can, except it does it at machine speed and across thousands of cases. Colorado’s law treats that risk as serious enough to require documentation, disclosures, impact assessments, and reports to the state attorney general. (coag.gov, leg.colorado.gov) xAI’s complaint says those rules are so broad they would reach a general-purpose chatbot like Grok if someone in Colorado used it inside a covered decision. The filing argues that forces one model to satisfy one state’s political and legal judgments even when the model is offered nationwide. (courthousenews.com) That is why the lawsuit is framed as a First Amendment fight instead of a normal compliance dispute. xAI says Colorado is compelling speech by requiring developers to “embed” the state’s views on topics like discrimination into the design and outputs of an artificial intelligence system. (bloomberg.com, courthousenews.com) Colorado wrote the law after a wave of concern that automated hiring and lending tools can quietly reproduce old human biases through new software. The statute gives the state attorney general exclusive enforcement power rather than letting private plaintiffs sue directly under this law. (leg.colorado.gov, regulations.ai) Even Colorado’s own leaders signaled discomfort with the bill right after it passed. Governor Jared Polis, Attorney General Phil Weiser, and state senator Robert Rodriguez said in 2024 that they wanted revisions before the law took effect, and the legislature later pushed the effective date back to June 30, 2026. (courthousenews.com, bhfs.com) That delay did not settle the argument. Consumer advocates kept defending Colorado’s approach as a needed guardrail, while technology companies and trade groups kept warning that fifty different state rules would turn artificial intelligence compliance into a patchwork map. (bhfs.com, reuters.com) The case matters beyond Musk because Colorado moved first where Congress still has not produced a single national framework for high-risk artificial intelligence. If xAI wins, states may have less room to regulate model design; if Colorado wins, other states get a roadmap for writing similar laws. (reuters.com, leg.colorado.gov) For now, the immediate question is narrower and more practical: whether a federal judge will let Colorado start enforcement on June 30 while the case is pending. xAI is asking for a declaration that the law is unconstitutional and an injunction blocking the state from enforcing it. (reuters.com, courthousenews.com)