California's Legislative Push on AI Employment Decisions
California lawmakers recently sent Governor Gavin Newsom a package of eight privacy bills and 16 AI bills, including SB 947, the "No Robo Bosses Act." This bill aims to prevent employers from relying solely on automated decision systems to discipline or terminate workers. If signed, it would require human corroboration for any AI-driven employment action, with a notice to the affected employee afterward.
This isn't about a single wrongful termination case. It's about recognizing that AI systems are already making employment decisions at scale, and existing labor protections weren't designed with algorithmic management in mind.
Key Dates and Implications
August 31, 2026: The California legislature sent the AI and privacy package to the governor.
September 30, 2026: Governor Newsom's deadline to sign or veto the bills.
If SB 947 becomes law, HR teams will face a new compliance requirement affecting performance management, discipline procedures, and termination workflows. The law won't ban AI in employment decisions but will prohibit unreviewed AI employment decisions.
Missing Controls in AI Employment Decisions
The issue isn't a single event but a lack of a control framework that matches how AI is used in HR systems.
Human Accountability Checkpoints Needed: Many AI-powered platforms can flag performance issues or generate termination paperwork quickly. However, they often lack a mandatory human review step before taking adverse action. Without this, an algorithm's recommendation can become a decision without human oversight.
Post-Decision Transparency Lacking: Employees often don't know an algorithm was involved in their discipline or termination. This lack of transparency can lead to fairness issues and legal risks if the decision is later challenged.
Cross-Functional AI Governance Missing: AI tools are often purchased and used by individual departments without a central function tracking which systems affect jobs, pay, or working conditions.
What SB 947 Requires
California's proposed SB 947 mandates that if an employer primarily relies on an automated system for discipline or termination, a human must corroborate the decision. The bill also requires notifying the employee post-use.
"Primarily relies" indicates the law applies when AI makes the decision, not just assists it. Human corroboration means someone with authority must review the AI's recommendation and ensure it's appropriate given the context. Employees must be informed if AI was involved in the decision-making process.
These requirements align with global AI governance frameworks, emphasizing meaningful human oversight. EU AI Act and ISO 37301 Compliance Management Systems include similar principles.
Action Items for Your Team
Map AI's Role in Employment Decisions: Conduct an inventory of systems influencing hiring, scheduling, performance evaluation, discipline, or termination. Document what each system does, the data it uses, and who reviews its output.
Integrate Human Review in Workflows: Ensure managers review data before acting on AI recommendations. This review must be genuine, not just a formality.
Train Managers on Corroboration: Managers should verify that decisions are fair and consistent with company policy. They need access to the data and explanations of AI recommendations.
Document Human Decisions: Keep records of who reviewed AI recommendations and what information they considered. This documentation is crucial if decisions are challenged.
Explain AI Systems to Employees: Prepare clear explanations of your AI systems and their role in decisions. Avoid technical jargon.
Review Vendor Contracts: Ensure your contracts provide the transparency and control needed for compliance. If a vendor can't meet these needs, consider other options.
California's AI legislation isn't final, but the principles behind SB 947 are already defensible under current labor laws. Start evaluating whether your AI systems are making decisions that should involve human judgment.
