AI news ·
California governor signs law requiring human review of AI-led workplace discipline
California bars employers from using automated systems as the sole basis for firing or disciplining workers, with a $500 penalty per violation starting July 1, 2027.
California Governor Gavin Newsom signed the No Robo Bosses Act into law on September 30, 2026, barring employers from using automated systems as the sole basis for disciplining or firing workers. Senate Bill 947, which takes effect July 1, 2027, requires human corroboration whenever automated output is the primary reason for an adverse employment action, creating new compliance obligations for HR, legal and operations teams across the state.
The law adds rules for workplace automated decision systems to the California Labor Code. It follows Newsom's 2025 veto of a broader predecessor and establishes requirements for employee notices, access to information and enforcement. The labor commissioner and public prosecutors handle enforcement, with violations carrying a $500 civil penalty each.
What the law covers
The act's definition of an automated decision system reaches beyond generative AI tools. It covers computational processes using machine learning, statistical modeling, data analytics or AI to produce scores, classifications or recommendations that assist or replace human judgment and materially affect people. Productivity ratings, attendance analysis and performance-tracking systems could fall within its scope, depending on how employers use them. Ordinary software infrastructure - spam filters, firewalls, antivirus programs, calculators, databases and datasets - is expressly excluded.
Covered employers include public entities, labor contractors and farm labor contractors. The central requirement distinguishes between exclusive and primary reliance on automation. Employers cannot use an automated decision system as the sole basis for discipline or termination. When they rely primarily on a system's output, they must have a person corroborate the decision using underlying data or other relevant supporting information, which may include supervisory evaluations, personnel records, work product, peer reviews or witness interviews.
Notice and data access requirements
Employees disciplined or dismissed primarily on the basis of automated output can request a meaningful, objective description of their own data used by the system. Employers must anonymize personal information belonging to customers, colleagues and other individuals when providing that description. A separate written notice requirement applies when automation is the primary basis for discipline or termination. The notice must stand alone, use plain language and be delivered in the language ordinarily used for workplace communications through a simple method, such as email or a hyperlink.
The obligation does not extend to every personnel decision that involves an automated tool. The law also prohibits using automated systems to obstruct compliance with applicable laws. Employers cannot use the technology to predict or take adverse action against workers for exercising legal rights protected by state and federal labor law.
Enforcement and evidentiary shifts
Enforcement rests with the labor commissioner and public prosecutors rather than a new standalone private right of action. The labor commissioner may investigate, order temporary relief, issue citations and bring civil actions. Public prosecutors can pursue civil enforcement, including injunctive relief, punitive damages and reasonable legal fees and costs.
The law shifts the evidentiary burden once use of an automated system in a disciplinary or termination decision is established. Employers must then demonstrate either that they did not primarily rely on the system or that they satisfied the human-review and notice requirements. Whether employees can pursue violations through California's Private Attorneys General Act remains a potential litigation issue. Compliance with SB 947's notice rules satisfies substantially similar automated-decision notice requirements under other state laws, but that protection does not cover warehouse quotas or other automated working-condition standards, and it does not remove obligations under the California Privacy Protection Agency's automated decision-making regulations.
Why this matters for executives and HR leaders
Employers preparing for July 2027 will need to examine not only which tools they use, but how managers rely on their output when making disciplinary decisions. The distinction between "sole" and "primary" reliance means a quick managerial sign-off on an automated recommendation will not satisfy the corroboration requirement. Legal and HR teams should audit current performance management, attendance and productivity systems now to identify where automated scores or classifications influence personnel decisions. For organizations building internal expertise on these requirements, AI Regulatory Compliance Courses offer structured training on navigating automated decision rules. Policy and public-sector leaders may also benefit from AI Public Policy Courses that address the intersection of automation and workforce regulation.