Disruption

Colorado senator defends landmark AI law, open to refinements before 2026

Senate Majority Leader Robert Rodriguez’s first-in-the-nation legislation was signed last week.
Colorado Senate Majority Leader Robert Rodriguez (D). (Courtesy of Colorado Senate Democrats)

Colorado Senate Majority Leader Robert Rodriguez (D) says he is open to “tweaking” his first-in-the-nation comprehensive artificial intelligence legislation that was signed into law last week and that he thinks it can serve as a starting point for other states.

“I commit to get this right,” Rodriguez said Thursday during a Pluribus Spotlight event on the landmark legislation, which takes effect in 2026.

Rodriguez said the new law, which Gov. Jared Polis (D) signed “with reservations” Friday, is a “starting point” as states begin to regulate the fast-moving technology. Rodriguez said he has already received “outreach” from the White House and from legislators in other states.

“I’m here as one person in one state starting a conversation that I hope my colleagues across the country can dig into more and improve and strengthen,” Rodriguez said.

The Colorado law focuses on high-risk AI systems that play a significant role in making consequential decisions about people’s lives in areas such as housing, lending and health care. It requires AI developers and deployers to take steps to ensure those systems do not discriminate and it requires disclosure to consumers so that they know when AI is involved. The law also creates a mechanism for consumers to challenge adverse decisions made by AI.

“This stuff is really cool, and the things it can do are great and awesome,” Rodriguez said. “But anytime you build or develop a product, there needs to be some accountability and responsibility with it.”

The Colorado law borrows heavily from legislation Connecticut Sen. James Maroney (D) introduced this year. That bill died in the House after Gov. Ned Lamont (D) threatened a veto.

Rodriguez said he first got to know Maroney after Colorado became the third state in the country to pass a comprehensive data privacy law in 2021. Maroney contacted him to learn more about the privacy law and the process for passing it. Maroney subsequently shepherded his own data privacy measure to passage in 2022.

Last fall, Maroney, Rodriguez and an informal working group of Democratic and Republican lawmakers from across the country started meeting to learn more about AI and come up with common definitions and ideas for legislation. Maroney’s bill grew out of that effort and a separate in-state task force process. Maroney shared the template with Rodriguez, as well as with colleagues in Virginia and Vermont who also introduced versions of the bill this year.

“This bill was always designed to be a floor or a chassis of artificial intelligence,” Rodriguez said. “[It] was never intended to be the end all, be all.”

Still, the state-led effort to pioneer AI regulations engendered pushback from some industry groups concerned about a patchwork of state laws. Consumer groups were concerned the guardrails were not protective enough. Rodriguez said his goal was to strike a balance between consumer protections and innovation.

“Sometimes when both sides are feeling a bit of pain, you’re in a good place,” he told the Pluribus News virtual audience, which included lawmakers from at least a dozen states.

Rodriguez acknowledged Polis’s call to “significantly improve” the law before it takes effect in 2026. He said a Colorado AI task force convening soon and the attorney general’s office writing rules to implement the law will provide opportunities to “make it more workable.”

Rodriguez defended the decision to take the lead on AI regulation and not wait for Congress, saying it is incumbent upon states “to take this on to protect their constituents.”

“Hopefully we’re setting a framework and a groundwork for the federal government,” he said.