In 2024, Colorado became one of the first states to pass a bill regulating modern artificial intelligence (AI) technologies. Senate Bill 24-205, signed by Gov. Jared Polis (D), aimed to curb algorithmic discrimination in “high-risk” AI systems—those influencing decisions in areas like employment, housing, education, and healthcare. However, in his signing statement, he cautioned against a complex regulatory regime and urged a federal solution.
Polis’ opinion recently gained attention when, in response to the surge of state-related bills, the U.S. House Commerce Committee proposed a national moratorium on state-level regulation of artificial intelligence.
After reports indicated the governor supported the moratorium, Reason Foundation reached out to Polis for a statement. His office responded:
Polis has said all along that a strong national policy governing AI consumer protection that supersedes state law would be the best course of action, as a state-by–state patchwork creates a challenging regulatory environment, and would leave consumers worse off overall. He is supportive of pre-empting state law for several years to give Congress time to figure this out and create a true 50-state solution to smart AI protections for consumers while driving innovation.
Polis’ support for federal preemption comes on the heels of an effort within Colorado to revisit the scope of a different AI law that failed to pass. Senate Bill 318, introduced this year, would have made key amendments to the existing AI law: exempting small businesses with less than $10 million in investment and $5 million in revenue; delaying implementation of SB 24-205 to 2027; narrowing the definition of algorithmic discrimination; and removing certain reporting and compliance duties. It also would have excluded low-risk tools like spellcheckers and document sorters from being treated as high-risk AI. Despite bipartisan interest, the bill was shelved in May.
By 2026, Colorado startups have to worry about not taking “reasonably foreseeable risks” into account with any AI implementation. And startups must provide updated documentation for compliance to prove they have mitigated these risks when there is a “substantial modification” to any AI system they may use.
In a post for the American Enterprise Institute, Will Reinhart attempted to estimate the hourly costs for complying with a California law related to algorithmic decision-making similar to Colorado’s SB 205 (but for larger businesses). He estimates that compliance costs will be around $20,000 per company for the first year. Now, imagine throwing more states’ AI regulations into the mix. Companies would have to expend significant resources to not only track hundreds of bills around the country but also navigate potentially subtle differences between each state.
As Polis argued in his 2024 signing letter, “[T]he important work of protecting consumers from discrimination and other unintended consequences of nascent AI technologies is better considered and applied by the federal government to limit and preempt varied compliance burdens on innovators.”