xAI sues Colorado
Elon Musk’s xAI has filed a lawsuit to block Colorado’s new AI law, arguing the state’s rules improperly restrict chatbot output and violate constitutional protections. The law—set to take effect in June—would impose anti‑discrimination safeguards and operational duties on AI systems used in high‑impact decisions, which xAI says interfere with product design. The outcome could set a national precedent on whether states can regulate AI behaviour before federal rules settle the field. (bloomberg.com)
Colorado wrote one of the first state laws aimed at artificial intelligence before Congress wrote a national one, and Elon Musk’s company xAI went to federal court on April 9 to try to stop it before the rules start on June 30. (reuters.com) The fight is not over chatbots in general. Colorado’s law targets “high-risk” artificial intelligence systems used to help make big life decisions like hiring, housing, lending, insurance, health care, education, and government services. (leg.colorado.gov) Colorado’s basic idea is simple: if a company builds or uses software that can tilt those decisions against people because of race, disability, religion, sex, or other protected traits, the state wants that risk checked before harm spreads at machine speed. The law calls that “algorithmic discrimination.” (coag.gov) The law puts duties on two groups. Developers have to use “reasonable care,” share documentation with customers, and tell the Colorado attorney general and known users within 90 days if they discover a credible discrimination risk in a covered system. (leg.colorado.gov) Companies that actually use the systems in Colorado also get homework. They must do impact assessments, keep risk-management policies, and notify consumers when a high-risk system is making or substantially helping make a consequential decision about them. (leg.colorado.gov) Governor Jared Polis signed the bill on May 17, 2024, but he also said he hoped lawmakers would refine it before it took effect. In March 2026, a Colorado work group backed a revised policy framework after months of complaints from business groups and technology companies that the original text was too broad. (coag.gov) (governor.colorado.gov) xAI says even that state-level approach goes too far. In its complaint, the company argues the law can reach model outputs and product design choices, which xAI says turns Colorado into a speech regulator for systems like Grok. (docs.reclaimthenet.org) (reuters.com) The company is not making just one constitutional argument. Reports on the filing say xAI is attacking the law under the First Amendment, the dormant Commerce Clause, due process, and equal protection, which is a way of telling the court the statute is vague, burdens interstate business, and treats speakers unfairly. (ppc.land) (reuters.com) That matters because Colorado is trying to regulate artificial intelligence the way states once regulated privacy and consumer finance first: start local, force companies to adapt, and let the rest of the country decide later whether to copy it. If xAI wins early, other states may think twice before writing their own rules for model behavior and high-risk systems. (reuters.com) (fpf.org) If Colorado wins, the state gets a green light to test a new kind of consumer-protection law before Washington settles the field. That would leave artificial intelligence companies facing the same map they already see in privacy law: one federal debate, fifty state experiments, and a lot of compliance teams in the middle. (reuters.com)
