California Executive Order Expands AI Oversight Through State Procurement

California Governor Gavin Newsom has issued Executive Order N-5-26, directing state agencies to develop new standards governing the procurement and use of artificial intelligence (AI). The order seeks to further regulate AI developers and deployers by laying the groundwork to impose additional guardrails for AI companies seeking to do business with the state. Such requirements will include enhanced certification, disclosure, and risk management expectations to be embedded in public contracting requirements.

CALIFORNIA’S PROCUREMENT-FOCUSED AI FRAMEWORK

Rather than imposing generally applicable legal obligations, the order directs state agencies to incorporate AI-related safeguards into contracting processes. In particular, the state is expected to develop vendor certification standards requiring companies to demonstrate that their AI systems incorporate controls addressing harmful or unlawful content, algorithmic bias, and impacts on civil rights and civil liberties.

These requirements, once implemented, could function as de facto regulatory obligations for companies seeking to access the California public sector market. The order also contemplates additional governance measures, including the potential use of watermarking for synthetic content and broader internal oversight of AI deployment within state government.

In addition to establishing a certification framework, the order authorizes California agencies to take an independent approach to supply chain risk. Notably, the state may assess federal determinations regarding restricted or high-risk vendors and in some circumstances proceed with procurement, notwithstanding federal restrictions. This decision appears to be a reaction to the recent determination by the federal government to label Anthropic a supply chain risk, which remains subject to ongoing litigation.

This aspect of the order introduces a first potential avenue for tension with federal national security frameworks and may create uncertainty for companies that operate across both state and federal markets. For instance, it is possible that such policy could be subject to constitutional challenges on the basis of preemption, especially if the state makes a different supply chain determination for a contract involving federal funds.

INTERACTION WITH FEDERAL AI POLICY

The order comes at a time of significant federal activity in the AI space. Recent White House initiatives, including executive actions and the administration’s AI Action Plan, reflect an increasing emphasis on national coordination, centralized oversight, and the development of uniform standards governing AI risks. At the same time, the federal government has been no stranger to leveraging government procurement to regulate AI.

Such strategy dates back to the first Trump administration, which published Executive Order 13960, Promoting the Use of Trustworthy Artificial Intelligence in the Federal Government, on December 3, 2020, outlining the initial framework for AI procurement by federal government agencies. The current administration published OMB Memo M-25-22 on April 3, 2025, which further detailed the requirements for AI procurement. Executive Order 14319, Preventing “Woke AI” in the Federal Government, mandated that federal agencies only procure large language models (LLMs) that follow “Unbiased AI Principles.” This executive order was followed by OMB Memo M-26-04, Increasing Public Trust in Artificial Intelligence Through Unbiased AI Principles, published on December 11, 2025, which described the specific requirements for LLMs to meet the “Unbiased AI Principles” requirements.

Both state and federal efforts aim to share a common focus on AI safety and bias mitigation, and both increasingly rely on procurement as a tool to influence private sector behavior. However, the federal government is increasingly seeing AI regulation as within its sole purview.

Among the items discussed in Executive Order 14365, Ensuring a National Policy Framework for Artificial Intelligence, published on December 11, 2025, the executive order directed the Department of Justice to challenge state AI laws deemed “onerous” and threatened to withhold federal Broadband Equity Access and Deployment Program funding from states with restrictive AI regulations. [1]

These terms could lay the groundwork for future litigation between state and federal governments over AI regulation, which will likely involve constitutional, statutory, and regulatory arguments.

IMPLICATIONS FOR GOVERNMENT CONTRACTORS

For government contractors, the California order could have significant near-term effects. Companies that contract with California should anticipate that AI-related representations, certifications, and compliance obligations will become a standard feature of state procurements.

These requirements are likely to focus on demonstrable risk management practices, including governance frameworks addressing bias, safety, and transparency. Even before formal standards are finalized, the California order provides a clear signal that the state will expect contractors to maintain robust AI compliance programs.

Federal contractors face a parallel but distinct set of considerations. Federal agencies are continuing to develop AI-related procurement requirements and enforcement priorities, and companies must be prepared to align with evolving federal expectations. For organizations that operate across both state and federal markets, the potential for divergence between the two regimes presents a particular challenge. Differences in certification requirements, as well as the possibility that California could depart from federal supply chain determinations, may create conflicting obligations and increased compliance complexity.

Addressing different standards between federal and state procurement requirements is not new for many contractors. For example, since early 2025 contractors have grappled with how to reconcile different requirements related to employment, as the federal government has prohibited affirmative action and DEI-related practices, while certain states have continued to require them in their contracts. While the subject matter is different, lessons learned in balancing these requirements are likely to be helpful here.

WHAT THIS MEANS

In light of these developments, organizations developing, deploying, or using AI technologies—particularly those engaged in public sector contracting—should take steps now to assess their governance frameworks against emerging expectations. This assessment includes evaluating internal controls related to bias mitigation and safety as well as preparing for enhanced contractual obligations tied to public sector procurement.

Companies should also closely monitor federal developments, particularly with respect to potential preemption of the state laws and regulations, as the balance between state and federal authority in AI regulation continues to evolve.

While California agencies are expected to issue implementing guidance in the coming months, which will provide greater clarity regarding the scope and application of the state’s certification framework, federal efforts to establish a national approach to AI governance are likely to continue, setting the stage for ongoing interaction—and potential conflict—between state and federal regulatory models.

HOW WE CAN HELP

Morgan Lewis is monitoring developments on the state’s recommended guidance and is available to counsel you or your business on the potential impacts of any proposals.