California Executive Order Focuses on Worker Protections for Disruption Caused by AI

Amid a wave of AI-adjacent layoffs, California Governor Gavin Newsom’s Executive Order N-6-26 (the “Order”) takes steps to protect workers from the potential disruption posed by ongoing AI innovation. According to the accompanying press release, it is part of “California’s comprehensive approach in creating commonsense guardrails balanced with opportunities to advance innovation.”

The Order directs California agencies and other entities to perform various analyses of the effects of AI on the workforce, with an initial specified time horizon of three to six months, which may support later legislation and regulation.¹

What is the Order focused on?

The Order focuses on transparency (including notice to employees and reporting to government agencies) and mitigation obligations (such as required severance and upskilling). Specifically, the Order directs state agencies, academic institutions, and other public entities to monitor and explore various ways to mitigate the effects of job losses stemming from AI. Salient directives include:

• Recommendations for legislative change: The Labor and Workforce Development Agency (LWDA), in coordination with other state agencies and industry partners, is directed to review and provide recommendations on revisions to the California Worker Adjustment and Retraining Notification (CalWARN) Act in a manner that provides early warning data on emerging industry trends.
• Identifying labor trends: The Employment Development Department (EDD) must provide, twice a year through the end of 2027, a report of feedback from businesses about the role of technological adoption in hiring or workforce decisions. The EDD must also launch a dashboard showing AI’s impact on employment across various sectors using unemployment insurance data. The LWDA is directed to provide a review of the emerging body of academic research on the potential workforce impacts of technological shifts. 
• Recommendations for mitigation obligations: The LWDA is directed to explore severance policies, subsidized employment, and other methods for assisting unemployed workers, including finding opportunities for training and upskilling. California’s Health and Human Services Agency, in coordination with the LWDA, the Office of Data and Innovation (ODI) and other relevant departments and agencies, must leverage ODI’s online platform to help Californians more easily navigate government services.
• Studying alternative approaches: The Order also calls for studying alternative worker business ownership models, ways to support small business’s technology adoption and education, and how the collective bargaining process addresses new technologies.

How does the Order relate to labor considerations?

This executive order may be understood as framing AI-driven workforce disruption as a labor issue. It signals alignment with organized labor concerns – even in non-union settings – which could impact future legislation and regulation for all California employers. Specifically, the Order emphasizes “worker voice,” collective bargaining, and existing protections, explicitly noting that collective bargaining agreements often address technology adoption and directing further review of how unions are handling AI (which suggests an intent to normalize bargaining over AI impacts as a policy expectation.) 

This creates overlap with labor law in two ways: (1) it lays groundwork for expanding the scope of issues unions may demand to bargain over (automation decisions, job impacts, retraining, compensation), and (2) it signals potential state efforts to replicate or encourage union-style protections (consultation, consent, compensation) both in union and non-union settings. 

The Order does not, of course, affect the subjects of bargaining in the private sector. Most private-sector bargaining is still governed by the federal National Labor Relations Act (NLRA). When it applies, the NLRA exclusively sets the mandatory subjects of bargaining. States cannot add subjects to the NLRA’s list—either in union or nonunion workplaces.  However, the impact of automation decisions on workers, including such matters as job impacts, retraining, and compensation, is a mandatory subject of bargaining under the NLRA.

Still, while not binding, the Order reflects a broader strategic shift toward embedding workforce protections into AI adoption frameworks, with unions and collective bargaining models serving as key reference points for future policy development.

Are efforts to protect workers from AI disruption limited to the California Order?

This legislative session, bills that address worker displacement caused by AI have been introduced in 7 states (California, Illinois, Maryland, Minnesota, Missouri, New York and Oklahoma) and at the federal level. A common theme across these bills is that AI-related displacement should not happen quietly or suddenly. 

Although none of these bills has become law at this time, they are helpful for understanding how transparency expectations are evolving. For example, California SB 951 would require 90 days’ written notice to the worker, EDD, and state regarding the specific AI system used, job functions automated, and the employer’s justification for adopting the technology. New York AB 9533 would require similar notice, with additional requirements to disclose available retraining or redeployment options and the AI vendor involved. 

At the federal level, United States Senate Bill 3339 seeks to require tracking of AI adoption and AI-related layoffs through existing surveys and voluntary reporting. This illustrates a growing interest in understanding how AI is reshaping employment, hiring, and workforce structures that can serve as an early-warning infrastructure and improve forecasting. 

What is the anticipated impact on workforce restructuring decisions?

Commentary regarding the actual relationship AI has with workforce restructuring has started to emerge. A desire to avoid criticism may prompt employers to closely review announcements and communications about AI-related workforce structuring, including layoffs. 

Eventually, legislation on worker displacement may end up being triggered based on how workforce structuring is explained. For example, was a layoff the result of employee displacement rooted in automation, employee productivity concerns, an increased investment in AI (and corresponding decreased investment in other functions), or a general cost-cutting measure that would benefit the organization’s AI development? Distinguishing between these explanations could directly impact related compliance obligations.

What can employers do now?

In anticipation of shifting standards regarding AI-related workforce restructuring, employers can:

• Carefully craft related statements, recognizing that the relationship with AI may impact future compliance obligations;
• Consider strategies for upskilling, reskilling, and redeploying employees as part of broader AI adoption strategy; and
• Monitor proposed laws that would impact transparency and mitigation efforts.