xAI Sues Colorado

Colorado wrote one of the first broad state laws for artificial intelligence, and now Elon Musk’s company is trying to stop it before the law even starts. On April 9, xAI sued Colorado Attorney General Phil Weiser in federal court and asked a judge to block enforcement before the law’s June 30, 2026 start date. (reuters.com) (courthousenews.com) The law xAI is targeting is Senate Bill 24-205, signed by Governor Jared Polis in May 2024. It covers “high-risk” artificial intelligence systems used in decisions about jobs, housing, lending, health care, education, insurance, and government services. (leg.colorado.gov) (content.leg.colorado.gov) Colorado’s basic idea is simple: if a computer system helps decide whether you get hired, approved for a loan, or accepted for housing, the company using it has to watch for unfair bias. The statute tells developers and deployers to use “reasonable care” to prevent “algorithmic discrimination” against Colorado residents. (content.leg.colorado.gov) (leg.colorado.gov) The law does not regulate every chatbot answer or every image generator prompt. It is aimed at systems that are a “substantial factor” in making what Colorado calls a “consequential decision,” which means a decision that changes a person’s access to a major life opportunity or service. (content.leg.colorado.gov) To get a presumption that they acted carefully, developers have to provide documentation about how a high-risk system works and what discrimination risks are known or reasonably foreseeable. Deployers have to run risk-management programs, complete impact assessments, review the system at least once a year, notify consumers, and offer a way to appeal some adverse decisions with human review when technically feasible. (leg.colorado.gov) (content.leg.colorado.gov) xAI says those duties do more than police business conduct. In its complaint, the company argues the law forces developers to build Colorado’s views about discrimination into the outputs of systems like Grok, and that this turns software design into compelled speech under the First Amendment. (courthousenews.com) (bloomberg.com) That speech argument is the hinge of the case. If a judge sees the law as a consumer-protection rule for tools used in hiring or lending, Colorado has more room to defend it; if a judge sees the law as the state dictating what an artificial intelligence model may say, xAI has a stronger First Amendment claim. (courthousenews.com) (reuters.com) Colorado’s law has been controversial almost since it was signed. Governor Polis, Attorney General Weiser, and bill sponsor Robert Rodriguez said in 2024 that they wanted to revise it, and lawmakers later pushed the effective date back from February 1, 2026 to June 30, 2026 instead of rewriting the whole framework. (courthousenews.com) (bakerbotts.com) That delay matters because Colorado moved ahead while Washington did not. Congress still has no single national artificial intelligence law covering bias, documentation, notice, and appeals across sectors, so states have been writing their own rules one by one. (reuters.com) (govtech.com) xAI is trying to stop that state-by-state model before it spreads. Its complaint warns that a patchwork of state rules would raise compliance costs and force companies to redesign nationwide systems to satisfy the toughest local standard. (courthousenews.com) (reuters.com) If Colorado wins, other states will have a road map for regulating artificial intelligence in the same way states regulate privacy or consumer lending. If xAI wins, the fastest path for artificial intelligence companies may be to challenge state laws as speech restrictions and push the fight back to Congress and the federal courts. (reuters.com) (courthousenews.com)