On the same day the White House cancelled a federal AI executive order, California’s Governor signed one. The two decisions — one a withdrawal, one an advance — define the political fault line that will determine how the world’s fifth-largest economy governs the most consequential labour market transformation since industrialisation.
By Vladimir Tsakanyan, PhD · Center for Cyber Diplomacy and International Security · cybercenter.space
The timing was not coincidental. On the morning of May 21, 2026, California Governor Gavin Newsom signed Executive Order N-6-26, directing state agencies to study the workforce impacts of artificial intelligence and develop policy recommendations for workers, small businesses, and communities facing economic disruption from AI adoption. Hours later, in Washington, President Trump cancelled a planned signing ceremony for a federal AI executive order, citing concerns that governance measures could impede American competitive advantage over China.
Two executives. Two orders. One day. The divergence they represent is not merely a difference in policy approach. It is the opening of a structural fault line in American AI governance — between a federal posture that treats competitive deployment speed as the primary value and a state posture that treats worker protection as an equally legitimate competing interest — whose resolution will shape not only California’s labour market but the regulatory architecture that every jurisdiction watching this contest will ultimately reference.
California is home to thirty-two of the world’s top fifty generative AI companies. A quarter of the technology’s patents and conference papers originate within the state. What California decides about AI governance is, by the weight of its economic significance, a decision that carries international as well as national consequence.
What the Order Actually Does
Executive Order N-6-26, effective immediately, imposes no new obligations on private employers today. Labour attorneys reviewing the order were consistent in their assessment: the immediate practical implications for California businesses are limited. The order’s significance is architectural rather than operational — it establishes a framework, sets deadlines for agency recommendations, and signals the direction of legislative travel that employers will need to navigate in the months ahead.
The order’s primary mechanism is a series of directed agency studies, each with defined timelines. By August 19, 2026, the California Employment Development Department must launch a public dashboard tracking AI’s documented impact on employment across industries in real time. By October 15, 2026, the Labor and Workforce Development Agency must review how collective bargaining processes are incorporating AI, including how worker voice is being included in technology adoption decisions. By November 17, 2026 — notably after Election Day — the LWDA must review the California WARN Act and recommend revisions responsive to AI-driven workforce disruption, including potential updates to mass layoff notice requirements, severance standards, and early warning systems for automation-related displacement.
The order also directs state agencies to evaluate worker ownership models that would allow employees to share in the productivity gains generated by AI adoption, to expand access to employment insurance programmes, and to develop training and upskilling pathways for workers displaced by automation. Higher education institutions are explicitly directed toward greater emphasis on on-the-job training to develop employment for recent graduates — a signal that the state views the mismatch between academic preparation and AI-era employment requirements as a governance problem requiring direct intervention.
Analyst note
The November 17 deadline for WARN Act recommendations — positioned after the November election — is the order’s most politically deliberate timing decision. The California WARN Act currently requires employers to provide sixty days advance notice of mass layoffs affecting fifty or more workers. The question of whether AI-driven workforce reductions should trigger equivalent notice requirements has been the subject of active legislative debate through SB 951, currently pending in the Legislature. By directing the LWDA to complete its review after Election Day, the order creates a basis for arguing that the Legislature should defer action on SB 951 pending the agency’s analysis — a procedural manoeuvre that simultaneously advances the worker protection agenda and manages the political timing of its most controversial elements.
The Legislative Landscape Surrounding the Order
Executive Order N-6-26 does not exist in isolation. It is the most recent element of a California AI governance architecture that has been developing through executive action and legislation since 2023, and that intersects with a body of pending legislation whose combined effect would, if enacted, constitute one of the most comprehensive AI labour market regulatory frameworks in the world.
The most significant pending legislation is SB 947, known informally as the “No Robo Bosses” Act of 2026. The bill, which has passed the State Senate and is currently pending in the State Assembly, would prohibit employers from using Automated Decision Systems to perform certain employment functions — including disciplinary decisions, termination, and deactivation — without providing workers with a written notice of the system’s role and, on request, a copy of the twelve months of the worker’s own data that the system used to reach its decision. The bill’s passage through the Senate reflects a legislative majority that has determined the use of AI in employment decisions requires specific accountability mechanisms that do not currently exist.
The CPPA’s September 2025 approval of automated decision-making technology regulations under the CCPA/CPRA adds a further layer, requiring employer notice requirements for automated employment decisions effective January 1, 2027. Combined with the executive order’s WARN Act review and the pending SB 947, California employers are navigating an AI governance environment that is simultaneously aspirational in its current state and materially more constraining in its observable trajectory.
The order’s relationship to Governor Newsom’s earlier AI executive actions is also significant. Executive Order N-12-23, issued in 2023, established California’s initial framework for AI risk assessment. Executive Order N-5-26, issued in March 2026, strengthened civil rights and privacy standards in the state’s procurement of AI technology and required transparency, civil rights protections, and watermarking of AI-generated content in government contracts. N-6-26 completes what is now a three-order architecture: governance of AI in state procurement, civil rights protections in AI deployment, and now workforce protection in the face of AI-driven economic disruption.
The Political Positioning
The political calculus behind Executive Order N-6-26 is not difficult to read, and its transparency is itself analytically significant.
Governor Newsom has been one of the most prominent Democratic voices on AI governance nationally, navigating a complex position between the technology industry whose economic base is concentrated in California and the labour movement whose political infrastructure has been central to Democratic electoral success in the state. The executive order represents a deliberate attempt to hold both constituencies simultaneously: it explicitly does not impose new burdens on employers today, which addresses the industry’s primary concern, while establishing a clear trajectory toward worker protection measures that the labour movement has been seeking.
The California Federation of Labor Unions President’s response to the order captured the political tension precisely. She described it as providing “a little sliver of hope” in certain areas — specifically the language on collective bargaining — while identifying what she characterised as a fundamental flaw: the order’s foundational assumption that catastrophic job loss is an inevitable outcome rather than a contingent one. The criticism from the left of an executive order that is itself a left-of-centre policy intervention illustrates the difficulty of the political territory the Governor is navigating: a labour movement that views the order’s analytical frame as already conceding too much to the inevitability of AI-driven displacement, and a technology industry that views any regulatory movement in this direction as a potential constraint on the sector’s growth.
The national political positioning of the order is also explicit. Newsom’s signing of N-6-26 on the same day as the federal AI order cancellation creates an unavoidable contrast that serves the Governor’s national political profile: California as the jurisdiction that moves toward governance while Washington retreats from it. The framing in the Governor’s own press release — “California has never sat back and watched as the future happened to us” — is a political statement as much as a policy one, directed at a national audience as much as a California constituency.
Analyst note
The simultaneity of the two May 21 decisions — Newsom’s signing and Trump’s cancellation — creates a political dynamic that extends beyond the immediate policy content of either order. Federal preemption of state AI law is an explicit objective of the White House’s AI legislative framework, which has signalled a push for federal standards that would supersede state-level regulation. California’s aggressive AI governance posture, accumulated through three executive orders and a body of pending legislation, is precisely the kind of state-level regulatory architecture that federal preemption would target. The political conflict between California’s regulatory direction and the federal posture is not merely a difference of approach. It is a structural confrontation over which level of government has the authority to set the rules governing AI’s impact on the workforce — a confrontation whose resolution will determine the regulatory environment for AI deployment across the entire US economy.
The Employer Implications: Near-Term and Medium-Term
For California employers, particularly those in industries with significant AI adoption — technology, healthcare, logistics, financial services, retail — the order’s practical implications unfold across two distinct time horizons.
In the near term, the order imposes no new legal obligations and requires no immediate action. The studies it directs are internal to state agencies. The recommendations it mandates are not yet in existence. Employers who are currently using AI in hiring, performance management, workforce planning, or operational automation are not subject to new requirements as a result of N-6-26 itself.
The medium-term picture is materially different. The WARN Act review, scheduled for completion by mid-November 2026, represents the most consequential near-term policy development for employers contemplating AI-driven workforce restructuring. The current California WARN Act requires sixty days advance notice for mass layoffs of fifty or more employees. If the LWDA recommends — as labour attorneys are widely anticipating — that AI-driven workforce reductions be treated as triggering events with potentially expanded notice periods, earlier warning requirements, and more detailed disclosure obligations about the role of AI in the employment decision, the compliance architecture for technology-enabled workforce restructuring will change significantly.
The “most compelling aspect” characterisation given by one labour attorney reviewing the order — specifically the WARN Act implications — reflects the assessment of the employer community that the notice requirement question is the one with the most immediate operational and financial significance. A company planning a technology-driven workforce reduction must currently give sixty days notice to affected workers. An expanded regime could require longer notice periods, earlier disclosure of the role of automated systems in the decision, and potentially new rights for affected workers to challenge or seek review of AI-assisted termination decisions.
The pending SB 947 adds a further dimension. If the No Robo Bosses Act passes the Assembly, the combination of automated decision-making disclosure requirements, CPPA ADMT regulations effective January 2027, and potential WARN Act expansion would create a compliance framework that requires California employers to maintain detailed records of AI system inputs, document the role of automated systems in employment decisions, and provide affected workers with both advance notice and data access rights of a kind that most current AI deployment practices do not contemplate.
The National and International Significance
California’s AI workforce governance architecture carries significance that extends well beyond the state’s borders, for reasons that are both economic and political.
Economically, the scale of California’s AI industry means that governance decisions made in Sacramento affect companies that operate nationally and globally. A WARN Act expansion that requires sixty or ninety days notice for AI-driven workforce reductions does not merely affect California workers. It affects the planning timelines of companies with California operations that are making workforce decisions with national scope. The compliance architecture for multistate employers must account for California requirements regardless of where the majority of affected workers are located, because California’s legal reach extends to the employment decisions of companies doing business within the state.
Internationally, California’s position as the home base of the majority of the world’s leading AI companies gives its governance decisions a standard-setting function that no other sub-national jurisdiction can claim. The EU AI Act’s approach to AI in employment — which imposes requirements for human oversight, transparency, and worker rights in high-risk AI applications — is developing in an environment where California’s regulatory direction provides a reference point for what a major AI-producing jurisdiction is prepared to mandate. The DLA Piper assessment that California’s order could have significant impacts on AI policy positions not only in California but in the United States more broadly is an understatement: California’s decisions in this domain are observed in Brussels, in London, and in Tokyo with the attention appropriate to the jurisdiction that houses the technology they are attempting to govern.
The broader geopolitical dimension is the one that the simultaneity of the two May 21 orders makes most visible. The federal government, on May 21, chose competitive deployment speed over governance constraint. California, on May 21, chose governance architecture over unconstrained deployment. The question of which approach ultimately prevails — through federal preemption, through the political success of the California model, or through some negotiated accommodation between them — will determine the regulatory environment within which the AI economy’s impact on the workforce is managed for the foreseeable future.
Bottom Line Assessment
Executive Order N-6-26 is, in its immediate operational content, a study directive. In its political and policy significance, it is the most comprehensive statement of governing philosophy about AI’s relationship to the workforce that any major jurisdiction in the United States has produced, and it arrives at a moment when the federal government has explicitly retreated from that territory.
The order does not resolve the fundamental tension at the centre of AI workforce governance — between the productivity and economic growth that AI adoption generates and the displacement and inequality that its unmanaged deployment produces. No executive order can resolve that tension. What N-6-26 does is establish California’s position in the negotiation over how that tension is managed: through worker ownership models, WARN Act expansion, collective bargaining protections, and the transparency requirements for automated employment decisions that the accompanying legislative agenda is pursuing.
Employers in California have been put on notice — not of immediate obligations, but of the direction of travel. The studies will be completed. The recommendations will be issued. The legislation will follow. The question for employers is not whether to prepare for the governance regime that N-6-26 signals, but whether to prepare now, while the recommendations are still being formed and the legislative process is still open to input, or later, when the obligations have been determined and the compliance window has closed.
California has, as its Governor said, never sat back and watched as the future happened to it. The future of AI workforce governance is being written now, in Sacramento as much as in Washington. Employers who understand that are better positioned than those who are waiting for the federal picture to clarify before engaging with the state one.
California · AI Governance · Newsom · WARN Act · SB 947 · No Robo Bosses Act · Workforce Disruption · AI Policy · Labour Law · Federal Preemption · Vladimir Tsakanyan
Leave a comment