Governor Gavin Newsom has fired back at the Trump administration over the federal government's recent move against a California-based artificial intelligence startup, issuing an executive order that asserts the state's authority to assess technology companies for itself. The order also introduces a new set of directives aimed at creating guardrails for the use of AI within state agencies, continuing California's push to lead on technology regulation.
The move comes in direct response to the U.S. Department of Defense's decision last month to label Anthropic, a prominent San Francisco AI developer, a "supply-chain risk." This designation effectively barred the company from bidding on certain military contracts. The federal action followed a disagreement over contract terms, where Anthropic sought to prohibit the U.S. military from using its AI systems for domestic mass surveillance or in the development of fully autonomous weaponry. While a federal judge has since issued a temporary injunction to block the designation, Newsom's order creates a new protocol for the state.
Under the new executive order, if a federal agency designates a business as a supply-chain risk, California will now conduct its own independent review before deciding whether state agencies can engage in business with that company. This establishes a potential conflict with federal directives and signals a willingness by the Newsom administration to chart its own course on technology governance, particularly when it impacts California-based firms. California is home to many of the world's largest AI companies and is the fourth-largest economy globally, giving its regulatory decisions significant weight.
New rules for state agencies
Beyond the direct challenge to the federal government, the broader purpose of Newsom’s executive order is to balance the promotion of AI innovation with the establishment of robust ethical safeguards. The order mandates that state agencies develop new standards for AI-related contracts to prevent the technology from being used to generate child sexual abuse material, violate civil liberties, or infringe upon legal protections against discrimination and unlawful surveillance. This represents a significant step in codifying responsible AI procurement at a state level.
The directive also requires state bodies to update the State Digital Strategy, identifying ways for generative AI to improve government transparency, enhance performance, and make public services more accessible. This includes developing user-friendly AI tools for Californians to interact with government services and providing "vetted GenAI tools" for state employees. Furthermore, the order calls for clear guidance on how government employees should watermark AI-generated images and videos to ensure transparency.
These measures are being implemented as California’s government is already expanding its use of AI. More than 20 state departments are currently involved in developing "Poppy," a generative AI assistant for government workers, while other agencies are testing AI for tasks like assisting state employees and providing resources for homeless individuals and businesses. This follows a previous executive order from 2023 aimed at studying and regulating generative AI, cementing the state’s proactive approach.
A tale of two administrations
The executive order draws a sharp contrast between Sacramento and Washington D.C. on AI policy. The governor's office explicitly criticized the Trump administration's approach, suggesting it has neglected the potential harms of the technology.
Unlike the Trump administration, California remains committed to ensuring that AI solutions adopted and deployed by [California]… cannot be misused by bad actors. At the federal level, President Trump has signed executive orders aimed at discouraging states from regulating AI and has pushed federal agencies to adopt the technology to accelerate decisions on Medicare and reduce federal regulations. Last month, the White House unveiled a national AI legislative framework that promotes a light-touch regulatory approach, notably lacking specific provisions to address issues of bias, discrimination, or civil rights. This divergence in philosophy echoes previous clashes between California and the Trump administration on issues like vehicle emissions standards and commercial rent control, as seen in the New York landlords oppose commercial rent control bill, where the state has consistently pursued more stringent regulations than the federal government.
Political pressures and a challenging road ahead
Newsom's navigation of AI policy is being closely watched from all sides, reflecting the complex political landscape in California. The state not only leads the nation in the volume of AI-related regulations but is also the epicentre of the industry itself. Tech giants like Google and Meta are major political donors, pouring money into influencing state politics ahead of the fall midterm elections.
Simultaneously, the governor faces pressure from organized labour. In February, union leaders pledged they would not support any potential presidential bid by Newsom without the implementation of stronger protections for workers against AI-driven displacement and disruption. This places the governor in a delicate position, balancing the interests of a vital economic sector with the concerns of a key political constituency.
This executive order is the latest example of Newsom using his authority to set a distinct policy agenda for California, positioning the state as a progressive counterweight to the federal government. This strategy has been a hallmark of his administration, which has previously issued orders on issues ranging from homelessness to public health. As both state and federal governments continue to grapple with the rapid advancement of artificial intelligence, California's actions will continue to have a major impact on the national conversation around technology and its role in society.
