Governor Gavin Newsom faces a critical decision window as the California legislature sends SB 947, the No Robo Bosses Act, to his desk. With only 25 days remaining until the September 30, 2026, deadline, the bill represents a significant attempt to codify human accountability in an era of increasing algorithmic management. Having passed the Senate with a 28-10 vote and the Assembly with a 53-14 margin, the legislation arrives with clear bipartisan support, yet its fate remains uncertain given the state’s complex history with similar regulatory efforts.
At its core, SB 947 seeks to establish the first state-level mandate in the U.S. requiring human oversight for AI-informed workplace termination and disciplinary actions. The bill explicitly bars employers from relying solely on automated decision systems (ADS) to fire or discipline staff. Instead, it mandates independent human verification whenever an ADS informs such a decision. Furthermore, the legislation requires employers to provide written notice to workers whenever an ADS has been utilized in a termination or disciplinary process. Beyond these procedural requirements, the bill categorically prohibits the use of ADS for predictive behavior analysis, the inference of protected characteristics, or taking adverse actions against workers for exercising their legal rights.
The current iteration of the bill is a direct response to the veto of its predecessor, SB 7, in October 2025. In his previous veto, Governor Newsom cited concerns regarding the bill’s overly broad scope and its potential to overlap with existing employment and discrimination laws. Proponents of SB 947 have revised the language to address these specific critiques, aiming to create a more targeted framework that complements, rather than conflicts with, established labor protections. Whether these adjustments are sufficient to secure the Governor’s signature remains the central question for enterprise builders and policy observers alike.
SB 947 does not exist in a vacuum; it is part of a broader, accelerating pattern of global and domestic AI governance. It draws clear parallels to the employment provisions within the EU AI Act, as well as NYC Local Law 144 and the Colorado ADMT Act. While NYC’s law focuses primarily on bias audits for hiring tools, California’s proposal pushes further into the realm of active management and termination, signaling a shift toward regulating the ongoing lifecycle of the worker-employer relationship.
The legislation also addresses the emerging challenge of agent governance. As AI agents increasingly assume supervisory roles, the question of who—or what—is actually making the decision to terminate an employee becomes a matter of legal liability. By requiring human oversight, SB 947 attempts to solve the agent delegation problem, ensuring that a human remains the final arbiter in consequential employment actions. This requirement is designed to prevent the ‘black box’ scenario where an autonomous system makes life-altering decisions without transparent, human-accountable reasoning.
If enacted, the law would take effect on July 1, 2027. Enforcement mechanisms are robust, involving the California Labor Commissioner, the California Attorney General, and local prosecutors. Crucially, the bill also provides for a private right of action, allowing workers to seek redress directly. Violations carry a $500 civil penalty per instance, alongside the potential for punitive damages and the recovery of attorney fees. For enterprises, the path forward involves preparing for a regulatory environment where the delegation of management tasks to AI is no longer a shield against accountability, but a process that must be documented, verified, and ultimately, human-led.