PubLedge

Developers and deployers of high-risk artificial intelligence systems would have been required to use reasonable care to protect consumers from known or reasonably foreseeable risks of algorithmic discrimination. The obligation was enacted in Colorado SB 24-205 but never became operative because the statute was superseded before its delayed effective date.

What Counts

• Developer controls addressing discrimination risks arising from intended and contracted uses
• Deployer controls addressing discrimination risks arising from deployment
• Risk management measures proportionate to known and reasonably foreseeable risks
• Compliance evidence sufficient to support the statute's rebuttable presumption of reasonable care

What Does Not Count

• Treating the enacted duty as currently operative after SB 24-205 was superseded
• Limiting review to intentional discrimination while ignoring foreseeable disparate impacts
• Relying on developer documentation without deployer-specific risk controls
• Generic responsible-AI principles without controls tied to the covered high-risk system

Implementing Legal Instruments