Colorado delays implementing first-in-the-nation AI law

Lawmakers in Colorado failed to reach an agreement on amending the state’s sweeping law regulating artificial intelligence, except to push back its implementation until June 2026.

Gov. Jared Polis had called a special session of the Colorado General Assembly earlier this month to vote on tweaks to the law, which he signed last year even while raising concerns about its effects on the technology industry. Polis also called the special session to discuss the budgetary impacts of Congressional Republicans’ “One Big Beautiful Bill” on the state.

But the session ended with no agreement on how the law should be amended, except that it should take effect on June 30, rather than the originally scheduled February. Lawmakers will likely try again to amend the law in January, when they return to Denver for their regular session. A bill delaying the law’s implementation passed both chambers easily.

“I believe this is the path forward to build on the progress we’ve made, resolve the outstanding areas of disagreement, and achieve a policy that protects Coloradans and businesses who utilize AI,” Senate Majority Leader Robert Rodriguez, who has been a leading sponsor on the state’s AI legislation, said on the floor.

The original law, which passed last year, set guardrails for companies that develop and use AI to try and protect Colorado residents from harm and discrimination. It puts various requirements on those who develop and deploy high-risk AI systems, like those that help make decisions on hiring, banking and housing. The law would require developers and deployers to avoid algorithmic bias or discrimination, and report any incidents to the attorney general’s office. It also placed reporting requirements on developers to consumers.

But the law came under fire from various technology companies, while Polis signed it reluctantly and urged amendments before it took effect.

And there looked to be an agreement on how to amend the bill during this special session. Rodriguez said there were the “outlines of a deal that would hold bad actors to account for their actions,” but it was torpedoed by Big Tech companies, as they objected to the liability requirements. Changes to the law would have made AI developers and deployers jointly liable for any discrimination consumers experience, even though developers do not interact directly with those consumers. It would also mean that developers would be responsible for shouldering some of the financial cost.

“Let me be clear here: we could be the first state in the nation to enact guardrails around AI that result in liability for these systems when they cause harm,” Rodriguez said.

The Colorado Chamber of Commerce called that provision “harmful” and urged lawmakers to vote against it. 

“The result would be higher legal risks for all businesses, fewer opportunities for AI development, and serious harm to Colorado’s business climate and competitiveness,” it warned this week. Rodriguez criticized opponents for killing the bill, and the overall deal he had struck.

“Overnight, the tech industry decided that they were so unhappy with the compromise that had been achieved by consumer protection organizations, educators, labor, and business that they would rather return to SB24-205!” he wrote. “At the end of the day, we must ask ourselves: why does this policy matter? What is at stake? Should a company whose AI system determines who gets hired or promoted, how much tenants pay for rent, or who receives medical care ever be held to account?”

After the delay passed both chambers, tech groups urged legislators to try again. Business Software Alliance Director of Policy Meghan Pensyl called the delay a “constructive step that affords stakeholders the opportunity to work together to get the state’s approach to AI policy right.”

“The proposed deal that had been circulated during the Colorado legislature’s special session fell short,” Pensyl continued in a statement. “Colorado now has an opportunity to craft clearer, more practical rules for high-risk AI that build public trust in technology and support AI adoption across the economy.”

When lawmakers return in January, the pressure will likely be on for them to make a deal in some form ahead of the bill taking effect. Proponents of the 2024 law said ahead of the special session that Colorado legislators should not lose sight of what is important.

"Colorado’s groundbreaking law provides a range of common-sense, baseline transparency and accountability requirements for the use of AI in consequential decisions such as housing and employment,” Travis Hall, state director at the nonprofit Center of Democracy and Technology, said in an email. “Coloradan workers, consumers, and citizens have made clear their support for guardrails when these tools are used in contexts that affect people’s lives and livelihoods. As the Colorado legislature convenes to consider revisions to this law, it must reject attempts to remove any meaningful oversight. We urge legislators to stand strong in their efforts to protect their constituents from the documented harms that the unaccountable use of these tools can cause."