Background
To address the risks in our increasingly digital society, Colorado has passed the Colorado AI Act, a landmark piece of legislation that addresses consumer protection from artificial intelligence systems.
The Colorado AI Act partially resembles the EU AI Act—a sweeping regulation protocol often deemed the strictest in the world—in its aim to uphold and protect consumer interests. One similarity may be observed in both laws’ recognition of “high-risk” AI systems. Such AI systems are those used to make “consequential decisions,” or significant decisions in the sectors of employment, education, healthcare, housing, and more. Another similarity between the two laws is the allocation of various obligations on developers and deployers of AI—companies that create AI systems and the businesses that use these AI systems in their work activities, respectively.
Key Provisions
Set to take effect in February 2026, the Colorado AI Act’s primary objective is to minimize algorithmic discrimination within “high-risk” AI systems. Algorithmic discrimination refers to conditions in which the use and output of an AI system results in unlawful treatment on the basis of ethnicity, sex, disability, age, or other factors.
In order to combat algorithmic discrimination, the Act imposes various duties upon AI developers and deployers. These include obligations such as disclosing and documenting the data used to train AI systems, documenting intended purposes and use cases of high-risk AI systems, and providing notices to consumers about the potential harms associated with high-risk AI systems. These requirements promote consumer protection, which advocates argue benefits Colorado residents and upholds novel consumer rights.
The Act also requires deployers and developers of AI to provide notice to consumers that AI is being used, ensuring that users are aware when AI is impacting decision-making. Additional consumer rights include the right to receive explanations of a high-risk AI system’s data and sources, as well as the right to appeal the decisions outputted by AI systems.
Pros: Novel Consumer Rights in a Digital Era
In an age where as many as 83 percent of employers use algorithms in hiring processes, the Colorado AI Act and its provisions may provide timely and relevant safeguards for the rapid proliferation of AI systems. Proponents argue that the Act is a necessary intervention to protect consumers from biased and ambiguous uses of AI in sectors that impact their lives. Advocacy group Consumer Reports applauds the Act as “a sorely needed floor of protections” for Coloradans in high-stakes decisions like housing, insurance, or employment—areas where people are often “totally in the dark” about how AI influences outcomes.
Moreover, supporters of the bill say that it will help prevent discrimination against marginalized communities in sectors where they already face major barriers. In an open letter to Colorado lawmakers, thirteen Colorado-based advocacy groups argue the law’s required transparency and impact assessments will help reveal biased or discriminatory uses of AI that might otherwise go undetected—providing both accountability and deterrence, which are foundational for consumer trust and safety.
Cons: Compliance Difficulties and Stifled Innovation
The Colorado AI Act has received backlash from technology and AI startups, the U.S. Chamber of Commerce, and Colorado businesses on the basis of compliance difficulties. Small businesses and startups with 50 or fewer employees have some exemptions under the bill, but compliance demands of the Act may still fall disproportionately onto these smaller businesses given that these companies often have smaller legal departments.
Critics also contend that the bill’s requirements may interfere with and slow the development and deployment of AI for these small businesses, potentially deterring AI innovation in Colorado. They highlight that new requirements might cost the state’s small businesses hundreds of thousands of dollars in compliance-related matters and divert corporations’ resources toward compliance matters that would otherwise be invested in job creation opportunities. Such increased compliance-related financial burdens on companies may render their services unaffordable to consumers and clients, forcing businesses to relocate elsewhere to continue such services.
Conclusion
When analyzing the potential harms of AI systems, algorithmic discrimination in AI systems continues to pose a dominant concern. The Colorado AI Act establishes mechanisms of accountability, transparency, and risk mitigation to combat these digital threats, serving as a historic benchmark and effort of AI governance in the United States. As other states such as California attempt to pass similar legislation related to algorithmic discrimination, they may turn to the Colorado AI Act to analyze what is appropriate when crafting a bill that balances consumer protection with AI innovation.
Although the Colorado AI Act, or SB25-004, was expected to take effect across the state on February 1st, 2026, a late August special legislative session called by Colorado Governor Jared Polis delayed the bill’s implementation until June 30th, 2026. During the session, over 150 lobbyists were in attendance as four different bills were brought to the floor, with each taking a different approach to AI regulation in the state. None of the bills received enough support to pass, and so Colorado lawmakers settled on extending the original act’s implementation date.
In the meantime, Colorado lawmakers will have the regular lawmaking term (January to May) to make changes to the Colorado AI Act as well as contend with President Donald Trump’s Executive Order No.14365. Signed on December 11th, 2025, the executive order intends to block states from issuing their own AI regulations, raising questions about the future of the Colorado AI Act.