Colorado today, the next state law tomorrow - without rebuilding your programme each time.

Five state regimes shape the current US state AI compliance picture, with several more in active legislative motion:

• Colorado SB 26-189 (effective January 2027). The replacement law for the original Colorado AI Act (SB 24-205, repealed via SB 26-189). Focuses on Automated Decision-Making Technology (ADMT) - AI systems making consequential decisions about Coloradans (employment, lending, housing, education, healthcare). Obligations: deployer impact assessments, consumer notice, opt-out rights, algorithmic discrimination protections.

  
  

• California SB 53 (in force). Adds AI auditing requirements for high-risk AI systems used in consequential decisions, with annual reporting obligations and CA Civil Rights Department oversight.

  
  

• California AB 2013 (in force). Generative AI training-data transparency - covered generative AI developers must publish summaries of training data used. Targets foundation model providers + GenAI vendors selling into California.

  
  

• Texas TRAIGA (effective 2026). Texas Responsible AI Governance Act - broader AI governance obligations covering AI systems deployed by state agencies and certain private-sector deployments. Includes algorithmic transparency, consumer protections, and AI risk assessment requirements.

  
  

• Emerging landscape. Connecticut, Virginia, Utah, and others have introduced AI-related provisions through consumer protection law, employment law, or sector-specific regulations. The pattern is fragmentation - each state framing the problem differently, with limited cross-state coordination.