Colorado's AI Policy Work Group unanimously approved a proposed rewrite of the state's landmark 2024 AI law on March 17, 2026, which would substantially weaken its regulatory requirements.[2] The new framework, titled "Automated Decision Making Technology in Consequential Decisions," would replace mandatory bias audits and risk impact assessments with a streamlined transparency-and-notice model focused on disclosure, correction rights, and human review.[2][5] Governor Jared Polis immediately endorsed the proposal, signaling strong political support for the overhaul.[2]
Who's Involved
The Colorado AI Policy Work Group—convened by Governor Polis and including technology industry representatives, consumer advocates, and business groups—developed the proposal.[2] The framework would affect AI developers and deployers using systems to make consequential decisions in hiring, employment, education, housing, insurance, finance, healthcare, public benefits, and government services.[3] If passed by the legislature, the revised law would take effect January 1, 2027, rather than the previously delayed date of June 30, 2026.[2][5]
Context and Timeline
Colorado passed the nation's first comprehensive AI antidiscrimination law (SB 24-205) in 2024, originally scheduled for February 1, 2026 implementation.[2] Industry and business communities immediately pushed back, arguing the requirements were unworkable and would stifle innovation.[2] Failed revision attempts during the 2025 legislative cycle led lawmakers to postpone enforcement to June 30, 2026.[2] Governor Polis then convened the working group to find common ground between tech interests and consumer advocates.
Why It's Newsworthy
The proposal represents a significant retreat from one of the nation's most aggressive AI regulations.[2][5] While the original law imposed sweeping obligations including bias audits, risk assessments, and mandatory reporting of algorithmic discrimination to the Attorney General, the new framework eliminates these governance requirements entirely.[5] This shift mirrors privacy law approaches rather than comprehensive AI governance models like the EU AI Act, making it potentially the first major rollback of AI regulation in the U.S.[5]