California Sets AI Procurement Standards as Federal Protections Erode
Governor Gavin Newsom signed an executive order on March 30 requiring companies that sell AI systems to California to meet new safety and privacy standards. The move creates a separate state procurement process independent of federal contracting rules, positioning California against the Trump administration's approach to AI regulation.
The order directs the Government Operations Agency to develop vetting requirements for AI vendors. Companies must document how their systems prevent exploitation of illegal content, detect and correct bias, and protect civil rights and free speech.
What the Order Requires
State agencies will evaluate AI vendors based on attestations about their safeguards. The California Department of Technology must also create recommendations for watermarking AI-generated images and manipulated videos - the first statewide standard of its kind.
The state plans to expand its own use of generative AI to improve service delivery. One initiative will create an AI tool to help Californians find state programs and benefits based on life events like starting a business or seeking employment.
Workforce Impact and Public Input
Newsom announced the launch of Engaged California, a digital platform that will gather public input on how AI policy should address workforce disruption. The platform previously helped Los Angeles residents shape recovery after the 2025 firestorms and allowed state employees to inform government efficiency efforts.
The state is rolling out the first statewide engagement effort through the platform in the coming months.
California's AI Market Position
California hosts 33 of the world's top 50 privately held AI companies and leads with 25% of all AI patents and conference papers globally. The Bay Area captured 51% of all U.S. AI startup funding between Q3 2024 and Q2 2025 - more than five times New York's share.
In 2024, California accounted for 15.7% of all U.S. AI job postings, ahead of Texas (8.8%) and New York (5.8%). Three California companies - Google, Apple, and Nvidia - have each surpassed $3 trillion in valuation.
Existing Protections
California already has laws addressing specific AI risks:
- Guidelines for frontier AI model development and deployment
- Safeguards against catastrophic harms from advanced systems
- Child safety protocols and protections against self-harm content
- Bans on nonconsensual sexually explicit deepfakes with watermarking requirements
- Protection of performers' digital likenesses
- Restrictions on AI-generated robocall scams
Government employees responsible for procurement or AI policy may benefit from understanding how these standards apply to vendor selection. AI for Government resources can help clarify how responsible AI governance works in practice. Those developing policy frameworks may find AI Learning Path for Policy Makers relevant for analyzing governance models and data-driven decision-making.
Your membership also unlocks:
