Colorado Lawmakers Advance Major Rewrite of Landmark AI Law Narrowing Scope and Delaying Enforcement

--- headline: "Colorado Lawmakers Advance Major Rewrite of Landmark AI Law, Narrowing Scope and Delaying Enforcement" slug: colorado-ai-law-sb189-rewrite-narrow category: policy story_number: "13" date: 2026-05-04 ---

With fewer than two weeks left in the 2026 legislative session, Colorado Democrats are moving to gut and replace the nation's first comprehensive artificial intelligence regulation -- a dramatic retreat driven by industry opposition, a federal lawsuit, and the political reality that a law meant to protect consumers from algorithmic bias never took effect.

Senate Majority Leader Robert Rodriguez of Denver introduced Senate Bill 189 on Friday evening, a measure that would repeal and rewrite Senate Bill 24-205, the landmark 2024 law that sought to impose sweeping obligations on developers and deployers of "high-risk AI systems." The replacement bill narrows the regulatory framework significantly, swapping SB-205's broad mandate for a leaner regime focused on transparency and consumer rights rather than proactive discrimination prevention.

From "High-Risk AI" to "Automated Decision-Making Technology"

The most consequential change is definitional. SB-189 jettisons SB-205's references to "high-risk artificial intelligence systems" and "algorithmic discrimination" in favor of a new legal category: "automated decision-making technology," or ADMT. The shift is more than semantic. Where SB-205 imposed affirmative duties on companies to assess, document, and prevent discriminatory outcomes from AI systems, SB-189 scraps those obligations entirely.

Gone are the requirements that developers conduct impact assessments, maintain risk management programs, and disclose to deployers the "nature, source, and extent" of personal data being processed. Rodriguez himself acknowledged the scale of the retreat: "The existing law is more comprehensive," he told reporters. "This one is more of a notice bill."

What remains are two core consumer protections. Companies and organizations using ADMT for "consequential decisions" -- defined as those affecting employment, education, housing, lending, insurance, healthcare, and access to government services -- must notify consumers that AI is being used and must provide an opportunity to appeal. Those requirements, while narrower than SB-205's framework, would still make Colorado one of a handful of states with enforceable AI transparency mandates on the books.

A Three-Year Shield From Penalties

SB-189 also pushes the law's effective date from June 2026 to January 2027, buying companies an additional seven months to prepare. Perhaps more significantly, it includes a three-year "right to cure" provision that allows AI developers and deployers to resolve violations without facing civil penalties under the Colorado Consumer Protection Act. Violations covered include failure to notify consumers about AI-driven decisions and failure to provide an appeal process.

Rodriguez said he insisted on one key guardrail: a sunset clause on the right-to-cure window. "That provision should not exist indefinitely," he said, signaling that the grace period was a concession he made reluctantly to secure industry buy-in.

The xAI Lawsuit That Forced the Issue

The rewrite did not happen in a vacuum. In early April, Elon Musk's xAI filed a federal lawsuit seeking to block SB-205 before its June 30 effective date, arguing the law was unconstitutionally vague and would chill AI innovation. The Trump administration's Department of Justice intervened on xAI's behalf -- the federal government's first litigation move to limit state-level AI regulation.

On April 27, a Colorado magistrate judge ordered Attorney General Phil Weiser not to enforce SB-205 until the state finalized its rulemaking for the law. Weiser subsequently agreed to suspend all enforcement activity, including investigations, until two weeks after a court rules on xAI's preliminary injunction motion. The joint motion filed by xAI, the DOJ, and Weiser to pause the lawsuit cited the likelihood that state legislators would replace the law before its effective date -- a prediction SB-189 now fulfills.

Cautious Optimism, Quiet Frustration

The reaction from consumer advocates has been measured. The People for Responsible Technology coalition -- whose members include the ACLU of Colorado, AARP Colorado, the Colorado AFL-CIO, the Colorado Education Association, and the Colorado Cross Disabilities Coalition -- described itself as "cautiously optimistic."

"It provides a path to hold developers and businesses using AI accountable when the technology makes consequential decisions for everyday Coloradans," said Dennis Dougherty of the Colorado AFL-CIO, speaking on behalf of the coalition.

But the gap between SB-205's ambitions and SB-189's scope is hard to ignore. The original law attempted to build a comprehensive framework requiring companies to understand and mitigate the discriminatory potential of their AI systems before harm occurred. SB-189 instead waits for harm to happen and gives consumers the right to be told about it and to challenge it -- a fundamentally reactive model.

Rodriguez framed the compromise philosophically: "If everyone is unhappy, that likely means a good compromise has been reached. I think this still gives basic consumer protections and some transparency."

What Comes Next

SB-189 must clear both chambers before the session ends in mid-May, and its late introduction means the legislative clock is tight. If it passes, Colorado's AI regulatory landscape will shift from the most aggressive in the nation to something closer to the middle of the pack -- stronger than the handful of states with narrow AI disclosure laws, but far weaker than what SB-205 envisioned.

The bill's passage would also likely moot the xAI lawsuit, removing the immediate legal threat but leaving unresolved the broader constitutional questions about how far states can go in regulating artificial intelligence. Other states watching Colorado's experiment -- and there are many -- will take note of the retreat. The message from the Centennial State is sobering for AI regulation advocates: even a Democratic trifecta, in a state that pioneered the first comprehensive AI law, could not withstand the combined pressure of industry lobbying, federal litigation, and the sheer complexity of regulating a technology that evolves faster than legislatures can act.