Colorado Lawmakers Grapple with AI Regulation in Special Session

Colorado Dems split on AI regulations ahead of special session

Why it matters: Big Tech, including Amazon and Meta, and policymakers in other states are closely watching to see where Colorado sets the bar for its first-in-the-nation AI regulations. Catch up quick: The guardrails put into law in 2024 are set to take effect next February. They would prohibit discrimination by algorithms in high-risk areas -- such as health care, education, employment, housing, insurance and finance -- and require greater transparency when used. Yes, but: Colorado Gov. Jared Polis reluctantly signed the bill, but now he is joining tech and finance industry executives in expressing concern that the regulations could hamper innovation. The latest: The Democratic legislative majority at the Capitol is split on how to best hold AI creators and companies using the technology accountable, and in a rare scenario, plans to present rival bills. Zoom in: The original bill's sponsors drafted legislation that simplifies current rules, while still requiring consumer disclosures and holding developers and companies accountable. * "The bill simply says that if technology is used to make major decisions about your life, you should know about it," said Rep. Jennifer Bacon (D-Denver) at a Wednesday rally in support of the legislation. A competing bill, led by moderate Democrats, is drawing bipartisan support. It would allow the state attorney general to pursue litigation against AI creators and deployers under the state's existing consumer protection laws, but limit individual lawsuits. * Bill sponsor Rep. William Lindstedt (D-Broomfield) told the Denver Post the bill tries to "find a middle ground that protects consumers and also protects our innovative economy here in Colorado." Between the lines: It's unclear where Polis stands on the legislation. * In a recent interview, he told Axios Denver he wants "to prevent discrimination from any source [including AI]... but make sure that this is a law that works for consumers and for our state." The other side: Two Republican bills also are set for introduction during the special session. One would repeal the original law, while the other would narrow the regulations to AI systems that make employment or public safety decisions. Neither is expected to advance.

Colorado lawmakers look to repeal, replace controversial artificial intelligence law

Shaun Boyd is Your Political Reporter at CBS News Colorado. Share you story ideas with her by sending an email to [email protected] or [email protected]. A year after Colorado state lawmakers passed the most sweeping artificial intelligence law in the country, they are now racing to repeal and replace it. The law is set to take effect in February 2026 -- one month after the regular legislative session begins. So Gov. Jared Polis directed lawmakers to take it up in a special session. Then, he called to address a $750 million budget deficit. Artificial intelligence uses algorithms to make decisions about everything from health care and housing to lending to and legal matters, and education and employment. In an effort to prevent discrimination in those decisions, Democratic state Sen. Robert Rodriguez of Denver passed a first-in-the-nation law to regulate how AI is used. "When you're the first one through the glass, you're gonna get a little messed up; Rodriguez told CBS Colorado. Rodriguez has come under fire from not only developers of AI but users such as hospitals and higher education. The law requires them to disclose possible biases and how they're mitigating them, even though they didn't develop the technology. "Everything became onerous and overly complicated," Rodriguez said. With tech companies threatening to pull out of Colorado, Rodriguez went back to the drawing board and introduced a bill in the special session to repeal and replace the law. Democratic state Rep. William Lindstedt from Broomfield did the same. Both bills require companies to disclose when AI is being used, but Rodriguez's bill defines AI much more broadly and requires companies to tell people which characteristics AI used to reach its decision, and if the data is inaccurate fix it. "If we're a leader in innovation and technology, what's wrong with being a leader in safe innovation and technology," Rodriguez continued. "I don't understand why the two can't go together." Lindstedt says he also wants safety and transparency. "Fourteen percent of new job growth is related to AI right now," Lindstedt said. "We have to get this right." Lindstedt's bill uses the federal definition of AI and requires companies to disclose when they use the technology but not how an algorithm arrived at a decision. "It makes sure that these AI companies comply with existing Colorado statutes, so the Consumer Protection Act, the Colorado Anti-Discrimination Act," Lindstedt said. "It will also provide needed transparency for consumers, so it clearly distinguishes the role of deployers and the role of these businesses." Tech companies largely support Lindstedt's bill. But tech companies say Rodriguez's remains unworkable. Fiscal analysts also estimated the state would need to spend nearly $7 million a year to comply with his bill's requirements, so Rodriguez amended it to exempt all public entities from having to tell people which of their characteristics AI used to make a decision. Rodriguez says individuals could still file open records requests to get the data. "We're trying to be thoughtful about this bill," Rodriguez said. "I hear that every bill is going to end the world as know it."