Colorado's AI Law Retreat: From Bias Prevention to Disclosure

Colorado abandoned its pioneering approach to regulating artificial intelligence in May 2026, replacing a law that required companies to actively prevent algorithmic discrimination with one that simply mandates disclosure when automated systems influence consequential decisions.

The shift represents more than legislative fine-tuning. It signals a fundamental change in how states are approaching AI governance—and it may not be enough to shield Colorado from federal legal challenges that could invalidate state AI regulation nationwide.

From prevention to notification

The original Colorado Artificial Intelligence Act, passed in 2025, imposed direct obligations on companies deploying high-risk AI systems in hiring, lending, insurance, and healthcare. Developers had to establish risk management programs, conduct impact assessments, and take proactive steps to prevent discriminatory outcomes based on race, gender, age, or other protected characteristics.

The Automated Decision-Making Technology Act, signed May 14, 2026, eliminates those requirements. Instead, companies must notify consumers when automated technology contributes to adverse decisions—such as loan denials or job rejections—and explain the technology's role within 30 days. Consumers can request data corrections and human review, but companies face no obligation to audit systems for bias before deployment.

The practical difference is substantial. Under the original framework, a lender using AI would need internal processes to catch discriminatory patterns before they affected applicants. Under the new law, that lender must only disclose AI's involvement after denying a loan. The burden shifts from institutions to individuals.

Why it matters

Colorado's legislative retreat exposes the constitutional vulnerability of state AI regulation. By narrowing its law to disclosure requirements, the state may have inadvertently strengthened the federal government's First Amendment case against it. If the Trump administration succeeds in blocking Colorado's scaled-back law, dozens of other states with pending AI legislation will face the same constitutional barriers. The outcome could determine whether states retain any meaningful authority to regulate AI systems that affect employment, credit, housing, and healthcare—or whether a "minimally burdensome national policy framework," as outlined in Trump's December 2025 executive order, will preempt state action entirely.

Federal challenge intensifies

Elon Musk's xAI sued to block the original Colorado law in April 2026, and the Department of Justice intervened days later—the first time the federal government moved to invalidate state AI legislation. The DOJ made two constitutional arguments: that the law's antidiscrimination framework violated equal protection by forcing race- and sex-conscious decisions, and that disclosure requirements constituted compelled speech under the First Amendment.

Colorado's rewrite eliminated the equal protection issue by removing all algorithmic discrimination provisions. But the First Amendment challenge may now be easier to litigate. The new law consists almost entirely of mandated disclosures—exactly what xAI characterized as forced communication that violates companies' right not to speak.

A federal judge stayed enforcement of the original law in April 2026, and that stay applies to the replacement. Colorado's law remains blocked until at least 14 days after the court rules on xAI's preliminary injunction request, which cannot be filed until 28 days after the state completes rulemaking.

The DOJ's intervention followed directly from Executive Order 14365, which President Trump signed in December 2025. The order directed the Justice Department to challenge state AI laws conflicting with federal policy and instructed the Commerce Department to identify "onerous" state laws by March 2026. That evaluation has not yet been published, but if Colorado's new law appears on it, the state will serve as a test case for broader federal preemption of state AI regulation.

These details were first reported by The Conversation, based on analysis from researchers tracking state AI legislation through the U.S. State AI Policy Tracker at the University of Denver's Daniels College of Business.