California Gov. Gavin Newsom signed Executive Order N-6-26 on May 21. The executive order is aimed at preparing California workers, businesses, and communities for the labor-market disruptions expected from artificial intelligence (AI).

Purpose Behind the Executive Order

The executive order responds to rapid AI adoption in California and the potential for job losses due to AI automation.  While it does not impose immediate new compliance obligations on private employers, it does mandate a series of agency reviews, reports, and tools to inform future legislation and policy. Employers who chalk this development up to a distant policy debate will likely be caught flat-footed.

What the order sets in motion

• A WARN Act rebuilt for AI automation. The Labor and Workforce Development Agency must recommend revisions to the WARN statute to make it more responsive to AI-driven layoffs and restructurings. The order further directs agencies to consider required severance, equity, and job training. This represents a sharp departure from California's long-standing practice of leaving severance to private bargaining.

• Government data that cuts against employers. Two new resources, an analysis of AI's disproportionate impact on demographic groups and an Employment Development Department dashboard tracking sectoral effects, will give plaintiffs and regulators an evidentiary head start in discrimination claims.

• A wider net. The order further ties workforce disruption to California's civil rights and privacy regimes. 

Where the litigation goes next

The fight will not be over whether a machine made the call. It will be over how it did, what data the employer fed it, what the human reviewer actually decided, and whether any of it was written down. Good documentation becomes even more important.

Five moves to make now

1. Map your AI. Resume screeners, productivity trackers, scheduling apps, and generative tools in performance reviews can all qualify as automated decision systems.

2. Test before you're asked. Implement validation testing to show tools are job-related and free of adverse impact.

3. Build flexibility into RIFs. Anticipate tougher WARN and severance requirements before they exist.

4. Keep humans in charge. Ensure managers reach their own conclusions and record the business reasons.

5. Govern centrally. One inventory, one owner, one defensible compliance record.

The shift is already here. AI is no longer just a tool employers can use to assist with HR functions; it has become yet another “decision” that will be scrutinized by law. The only open question is how fast employers adjust.