On August 28, Colorado Gov. Jared Polis signed SB4, delaying the effective date of the Colorado AI Act (CAIA) from February 1, 2026, to June 30, 2026. Earlier this summer, Gov. Polis called the Colorado Legislature into a special session on the CAIA in an effort to reduce its fiscal impact on the state of Colorado, which is facing a $750 million budget shortfall.
This special session came shortly after the release of President Donald Trump’s AI Action Plan, which urges the federal government to avoid supporting states that impose burdensome AI regulations. While it respects states’ rights to pass reasonable laws, it recommends only providing federal AI-related funding to states whose regulatory environments do not stifle innovation.
Although Colorado lawmakers considered broader changes to the CAIA — such as reducing the scope of required disclosures by AI developers and deployers, curtailing certain consumer rights provisions and establishing a framework of joint and several liability for AI developers and deployers — they were unable to reach an agreement.
The CAIA regulates AI systems used to make or facilitate consequential decisions, such as those pertaining to healthcare, employment and essential government services, which it defines as “high-risk AI systems.” Developers and deployers of high-risk AI systems will be required to make certain disclosures about:
- How the AI systems are used
- How the AI systems were developed
- How risks of algorithmic discrimination are measured and mitigated
The CAIA also gives Colorado consumers certain protections, including:
- The right to be notified before a decision is made using a high-risk AI system
- The right to correct incorrect personal data used in AI decision-making
- The right to appeal adverse consequential decisions made by a high-risk AI system
The CAIA will be enforced by the Colorado attorney general. Violations will be treated as unfair trade practices, subject to a penalty of up to $20,000 per violation.
