Colorado Refines AI Act to Target Algorithmic Discrimination
The state is implementing a risk-based regulatory framework to prevent bias in high-risk automated decision systems.
Colorado is moving forward with a regulatory framework designed to curb algorithmic discrimination through the implementation of its revamped AI Act. The state's approach focuses specifically on high-risk artificial intelligence systems to ensure consumer protections are embedded in automated processes.
The state has transitioned to SB 26-189, which replaces the original AI Act (SB 24-205) to refine how the state manages automated decision-making technology (ADMT). This legislation targets high-risk AI systems used for consequential decisions, specifically focusing on critical sectors such as healthcare, housing, and hiring.
The Shift to Risk-Based Regulation
Colorado's legislative evolution reflects a broader trend across the United States, where individual states are establishing their own guardrails in the absence of a comprehensive federal AI law. By focusing on "high-risk" systems, Colorado is mirroring elements of the European Union's AI Act, which categorizes AI tools by their potential for harm rather than applying a blanket set of rules to all software.
This risk-based approach allows regulators to concentrate oversight on systems that have a direct, significant impact on a person's life opportunities. By refining the definitions of ADMT under SB 26-189, the state aims to create a more precise mechanism for identifying and mitigating bias before it affects the public.
Industry Implications
These rules are significant because they set a potential precedent for how other state regulators define and police "algorithmic discrimination." For companies operating within Colorado, the act introduces specific documentation and auditing requirements. Businesses deploying high-risk AI must now be prepared to prove their systems are not producing biased outcomes, shifting the burden of proof toward the developer and operator.
As more states look to Colorado's model, the industry may face a fragmented landscape of compliance requirements. Companies will likely need to implement rigorous internal governance and third-party audits to meet the state's standards for transparency and fairness.
Future Outlook
While the legislative framework is established, the industry is now watching for the finalization of the specific rules that will govern day-to-day enforcement. Key areas of focus include the exact metrics regulators will use to measure discrimination and the frequency of required audits. As the state transitions from the original act to the refined SB 26-189, the focus remains on whether these rules can effectively prevent bias without stifling technological innovation.