Establishes consumer protections and imposes disclosure, evaluation, and performance obligations on developers and deployers of high-risk AI systems that may generate algorithmic discrimination in making consequential decisions (related to employment, housing, health benefits, etc.).
Analysis summaries, actor details, and coverage mappings were LLM-classified and may contain errors.
This is a binding state statute enacted by the Colorado General Assembly with mandatory obligations, enforcement mechanisms through the Attorney General, and penalties for non-compliance as an unfair trade practice.
The document has good coverage of approximately 5-6 subdomains, with strong focus on unfair discrimination (1.1), unequal performance across groups (1.3), privacy compromise (2.1), lack of transparency (7.4), and lack of robustness (7.3). Coverage is concentrated in discrimination/toxicity and AI system safety domains, with minimal coverage of other risk areas.
The document governs AI use across multiple sectors through its definition of 'consequential decisions' which explicitly covers education, employment, financial services, government services, healthcare, housing, insurance, and legal services. The regulation applies horizontally across these sectors rather than being sector-specific.
The document covers multiple AI lifecycle stages with primary focus on deployment and operational monitoring. It addresses planning through risk management requirements, model development through documentation of training data and evaluation, verification through impact assessments, deployment through notification requirements, and ongoing monitoring through annual reviews and post-deployment oversight.
The document explicitly defines and covers AI systems broadly, with specific focus on high-risk AI systems that make or substantially influence consequential decisions. It does not mention frontier AI, general purpose AI, foundation models, generative AI, predictive AI, open-weight models, or compute thresholds. The scope is defined functionally based on use case (consequential decisions) rather than technical characteristics.
Colorado General Assembly; Senator Rodriguez; Senator Cutter; Senator Michaelson Jenet; Senator Priola; Senator Winter F.; Senator Fenberg; Representative Titone; Representative Rutinel; Representative Duran
The document is a Senate Bill proposed and enacted by the Colorado General Assembly, with specific senators and representatives listed as sponsors.
Colorado Attorney General
The Attorney General has exclusive authority to enforce this Act and can promulgate rules for implementation. Violations constitute unfair trade practices enforceable by the Attorney General.
Colorado Attorney General
The Attorney General has monitoring authority to require developers and deployers to disclose documentation, impact assessments, and risk management policies for evaluation of compliance.
The Act explicitly targets developers and deployers of high-risk AI systems doing business in Colorado. Developers are defined as persons who develop or intentionally and substantially modify AI systems, while deployers are persons who use high-risk AI systems.
6 subdomains (4 Good, 2 Minimal)