← Back to feed

California Governor Gavin Newsom signed an executive order on May 21 directing state agencies to prepare for AI-driven workforce disruption — the first executive action of its kind in the US to treat the economic effects of AI as a state policy problem rather than a private sector one.

The order arrived the same day Meta announced it was laying off approximately 8,000 workers, roughly 10% of its workforce, citing the acceleration of AI investment. It was also the same day President Donald Trump postponed a planned federal AI pre-release vetting executive order, saying “I didn’t like what I was seeing.”

What the order directs

The executive order mobilises California’s state agencies, labour economists, universities, and industry leaders around three goals: building early-warning data infrastructure, creating a policy safety net for displaced workers, and ensuring productivity gains flow to workers rather than exclusively to capital.

On data infrastructure:

• A new AI labour dashboard tracking sector-level displacement
• Business feedback on AI’s role in workforce decisions incorporated into California’s monthly jobs report
• Recommendations within 180 days on revisions to the California WARN Act — the law requiring advance notice of mass layoffs — to capture early signals of AI-driven displacement before official unemployment numbers register it

On worker protection:

• State review of severance standards and equity compensation as transition support
• Expanded unemployment insurance and employment stability payments
• An AI playbook to modernise job retraining programs, with updated target industries
• A single online platform consolidating access to government support services

On wealth distribution:

• Evaluation of worker ownership models and employee-owned company structures
• Development of “universal basic capital” concepts — giving residents a stake in corporate stocks, bonds, or state wealth funds generated by AI-driven productivity

The data problem the order is trying to solve

The executive order addresses a structural limitation of existing labour statistics: AI can eliminate portions of a job — specific tasks, subtasks, workflows — before it eliminates the entire role. When that happens, payroll and hiring data may look stable while the actual skill demand and workload composition are changing underneath.

By the time mass layoffs appear in WARN notices and unemployment filings, a sector may already be hollowed out. The order’s emphasis on monthly jobs report integration and revised WARN Act triggers is specifically aimed at moving the measurement window forward.

Political context

The AFL-CIO delivered an ultimatum to Newsom in February: regulate AI or lose labour’s support for any future presidential run. AFL-CIO President Liz Shuler called a potential AI economic collapse a coming “crisis.” The executive order is, in part, a direct response to that pressure.

Newsom has also been building towards this. In August 2025, he announced a partnership with Google, Microsoft, IBM, and Adobe to expand AI education in California schools. That effort, which focused on access to AI tools, now looks like a precursor to a more substantive policy framework that focuses on who captures the gains from those tools.

“California will pursue new policies that make sure working Californians — not just Big Tech — benefit from the wealth and breakthroughs coming out of this space,” Newsom wrote on May 21.

Federal-state tension

Trump’s simultaneous postponement of the federal AI vetting order matters for context. The administration has signalled it does not want states imposing substantive AI regulation — Trump’s December 2025 executive order attempted to curtail state-level AI laws, though it did not directly preempt them.

Whether California’s labour-focused executive order survives federal preemption challenges will depend on how the White House’s national AI legislative framework, released earlier this month, interacts with state labour law authority.

Key numbers

• 180 days: deadline for WARN Act revision recommendations
• 8,000: Meta layoffs announced the same day the order was signed
• ~40M: California workforce affected by the order’s scope
• August 2025: prior Newsom AI education partnership with Google, Microsoft, IBM, Adobe
• February 2026: AFL-CIO ultimatum to Newsom on AI regulation