California: SB 947 Mandates Human Oversight on AI Layoffs

California has adopted a framework that prohibits fully automated layoffs and mandates human review and explanation for employees. Starting July 1, 2027, a right to appeal to the Labor Commissioner will be established, and an obligation to indicate whether a wave of layoffs is caused by an AI system will be introduced. This advancement comes as the use of AI in human resources becomes widespread and other states are testing more or less strict approaches.

Effective Date, Appeals, and Disclosure of Layoff Plans Caused by AI

The law stemming from SB 947 will take effect on July 1, 2027. It allows affected employees to appeal to the California Labor Commissioner if they believe the prescribed procedure was not followed during a layoff decision. Employers will also be required to indicate whether a mass layoff, relocation, or termination is caused by an AI system, according to a statement published on the California governor's website. Such protections will remain fragmented from state to state until a federal version is adopted.

What SB 947 Prohibits, Controls, and Requires from Employers

It will be prohibited to fully delegate the responsibility for disciplinary decisions and layoffs to automated decision-making systems. It requires that a human verify each decision made by an automated system, and that an explanation of the reasons for that decision be provided to the employee. This justification must detail all data used, such as personnel records, evaluations, work outputs, peer reviews, and witness interviews. When the employer is unable to confirm the results of the system or if the person tasked with reviewing them believes they are inaccurate, incomplete, or misleading, these results cannot serve as the basis for a sanction or layoff. However, the law does not prohibit the use of automated systems in these processes.

AI Already Ubiquitous in HR Practices, According to a Survey

A survey conducted in April by MyPerfectResume among 1,000 HR professionals reports that 73% use AI to make hiring decisions and 52% to inform "restructuring and role planning." More than half, 51%, use it to flag candidates deemed at risk. Only 26% report not using AI to decide on layoffs. Beyond these figures, the use of AI is growing in daily tasks, recruitment, and performance evaluation, in a context where its role in hiring and contract termination decisions is increasing.

Outside California, Mandatory Disclosures and Laboratory Audits

Several states are experimenting with other levers. Colorado and Connecticut require the disclosure of AI use in hiring and workplace processes. Illinois has, this summer, adopted legislation becoming the first to require third-party audits of cutting-edge laboratories. California and New York are leaning towards stricter measures, while initial protections are emerging at the state level. In contrast, at the federal level, regulation remains less specific and relies on voluntary commitments. The California framework could serve as a model for other jurisdictions.

Risks of Evasion and Counterweights, According to Attorney Lily Li

Attorney Lily Li, founder of Metaverse Law, sees the California law as a precedent for other states, while anticipating attempts to circumvent the required human oversight. She believes some employers might settle for superficial compliance reviews to avoid the costs of a genuine human review. She considers the ability for employees to demand an explanation as a safeguard. However, it is possible that the burden of contesting unverified AI-driven layoffs will fall on plaintiffs in court.