California Will Independently Review Federal AI Supply-Chain Designations
California will no longer automatically accept federal decisions to bar companies from state contracts based on supply-chain risk designations. Gov. Gavin Newsom signed an executive order Monday requiring state agencies to review such designations independently and decide whether to do business with affected companies.
The order responds directly to a dispute between the Department of Defense and San Francisco-based AI company Anthropic. The DoD designated Anthropic a supply-chain risk last month after the startup refused contract terms that would have allowed the military to use its systems for domestic mass surveillance and fully autonomous weapons. A federal judge recently issued a temporary injunction blocking that designation.
Newsom's order addresses a broader tension: state agencies need guardrails on AI use while also accelerating adoption of the technology. California hosts more of the world's largest AI companies than any other state and leads the nation in AI regulations.
What State Agencies Must Do
- Develop contract standards for AI systems, with specific attention to risks involving child sexual abuse material, civil liberties violations, discrimination, unlawful detention, and surveillance
- Create a list of vetted generative AI tools for employee access
- Update the State Digital Strategy to show how generative AI can improve government transparency, performance, and public access to services
- Develop generative AI applications that help Californians access government services
- Issue guidance on watermarking AI-generated images and videos
More than 20 California departments are already working with or developing Poppy, a generative AI assistant for state employees. About half a dozen agencies are testing AI for tasks like supporting staff and assisting homeless people and businesses. State courts and local governments are also increasing their use of the technology.
Federal-State Divide on AI Policy
Newsom's office framed the order as a counterweight to federal policy under the Trump administration. The White House has signed executive orders discouraging states from regulating AI and has urged federal agencies to adopt AI to reduce regulations and accelerate decisions on programs like Medicare.
The Trump administration's AI policy framework, introduced last month, takes a minimal regulatory approach and does not address bias, discrimination, or civil rights concerns.
This is Newsom's second executive order on AI. A 2023 order focused specifically on generative AI, calling for more state agency use while requiring guardrails. Both orders reflect a strategy of state-level action where federal guidance has been sparse or permissive.
Political Pressure From Multiple Sides
Newsom's AI decisions face scrutiny from labor unions and technology donors. In February, union leaders said they would not support his presidential ambitions without stronger worker protections from AI. Technology companies, meanwhile, are investing heavily in California politics ahead of midterm elections this fall.
For government employees, the practical effect is immediate: your agencies will soon have official guidance on which AI tools are approved, how to use them safely, and what restrictions apply to their use. The standards being developed will shape what AI systems your workplace can and cannot adopt.
Your membership also unlocks:
