The amended law splits liability for algorithmic discrimination between developers and deployers based on relative fault, with enforcement remaining exclusively with the Colorado Attorney General rather than private plaintiffs. The specific contours of that liability allocation are not yet fully detailed in public filings.
Colorado's original AI statute, SB 24-205, passed in 2024 as the nation's first comprehensive state AI law with an initial February 1, 2026 effective date. That deadline shifted to June 30, 2026 following employer and vendor pushback. In 2026, a state AI policy work group proposed replacing the broader framework with the narrower automated decision-making model, stripping out bias-audit and risk-assessment requirements in favor of transparency, notice, and human-review obligations.
The revision matters because Colorado's approach will likely influence other states' AI governance strategies, particularly for employment contexts. The shift from broad system regulation to decision-specific rules reflects a national debate over whether AI regulation should cast a wide net or target the specific decisions where discrimination risk concentrates. HR teams and AI vendors should monitor the January 2027 effective date and prepare compliance infrastructure around notice, adverse-action procedures, and record retention now.