News California AI Bills Advance to Appropriations 2026

On April 22, 2026, several California artificial intelligence bills advanced from policy committees to the Assembly Appropriations Committee, moving to the fiscal phase of the legislative process.

Bills Advancing

AB 2027 (Ward) — Worker Data Protection

Assembly Bill 2027, authored by Assemblymember Matt Ward (D-San Diego), would restrict California local government employers from using worker data to train AI systems that could automate or replace jobs. The bill specifically prohibits the use of employee data for training AI tools when the goal is to eventually automate that job.

The bill passed the Assembly Labor and Employment Committee and now moves to Appropriations for fiscal review.

AB 2575 (Ortega) — AI in Health Care

Assembly Bill 2575, authored by Assemblymember Liz Ortega (D-San Leandro), establishes safeguards for AI and clinical decision support systems in health care settings. The bill:

• Prohibits AI developers from evading liability by claiming a health care worker's failure to override an AI output severs their liability for harm caused
• Mandates transparency notices requiring health facilities to inform users how AI tools work, their limitations, data sources, and known risks
• Protects workers' rights to override AI outputs based on their professional judgment without employer retaliation
• Requires patient disclaimers when generative AI is used for clinical communications

Hospital groups have opposed the bill as overly broad.

AB 2653 (Lee) — Sweatfree AI Code of Conduct

Assembly Bill 2653, authored by Assemblymember Alex Lee (D-San Jose), extends California's existing Sweatfree Code of Conduct to state contracts involving AI products that require data enrichment services. The bill:

• Mandates the Department of Industrial Relations to update its contractor responsibility program by July 1, 2027
• Requires bidders on relevant state contracts to certify no goods are produced by sweatshop labor, forced labor, or exploitative labor laundering in AI data enrichment supply chains
• Defines sweatshop labor as work violating minimum standards on wages, safe conditions, discrimination/harassment, and just cause termination
• Establishes an advisory panel (majority from trade unions and civil society) to help develop the code
• Requires independent monitoring of complaints by DIR, labor organizations, or qualifying nonprofits

The bill targets exploitative labor often found in global AI data labeling supply chains.

Earlier Committee Advancement: SB 1142

Senate Bill 1142, the Digital Dignity Act authored by Senator Josh Becker (D-Menlo Park), also advanced from committee earlier in April 2026. The bill:

• Strengthens penalties for defamation or false impersonation using AI-generated digital replicas
• Requires large online platforms creating and hosting generative AI content to provide mechanisms for individuals to remove unauthorized digital replicas
• Imposes enhanced liability including liquidated damages of $1,000 or actual damages for each violation by distributors with actual knowledge
• Requires generative AI providers to maintain records for at least 90 days to comply with court orders
• Enables enforcement by city attorneys or the Attorney General

The Senate Privacy, Digital Technologies, and Consumer Protection Committee advanced SB 1142 after debate over free speech and takedown rules.

Next Steps

All four bills now move to the Appropriations Committee, the final committee hurdle before potential floor votes. The Appropriations Committee will review the fiscal impact of each bill before determining whether they advance to the full Assembly (for AB measures) or Senate (for SB 1142).

See Also

• Legislation
• California
• State Legislation
• Employment
• Healthcare AI
• Deepfakes