Blogs

Can AI be the boss? California bill targets automated workplace decisions

9/1/26

By: Sunshine Fellows

Artificial intelligence is moving beyond recruiting and resume screening and into the day-to-day management of employees, including performance monitoring, scheduling, discipline, and even termination recommendations. California lawmakers are now considering legislation aimed directly at that development. Senate Bill 947 (“SB 947”), which has advanced through the California Legislature and is awaiting further action in the Assembly, would impose significant restrictions on employers’ use of automated decision systems in workplace decisions. The legislation reflects an emerging regulatory concern: as AI becomes increasingly capable of acting like a manager, lawmakers are asking how much authority employers should be permitted to delegate to an algorithm.

From AI assistant to “AI boss”

Employers already use automated systems to evaluate applicants, monitor productivity, assess performance, and identify workforce trends. SB 947 focuses on what happens when those systems begin influencing consequential decisions about existing employees. Among other restrictions, the bill would prohibit employers from relying solely on an automated decision system to make disciplinary or termination decisions. When an employer primarily relies on such a system to assist with discipline or termination, the legislation would require an independent investigation by a human reviewer to corroborate the decision.

The proposal would also impose transparency requirements. Under the current version of SB 947, an employee affected by certain AI-assisted disciplinary or termination decisions would receive a written post-use notice, and employees could request certain data used by an automated system in making decisions about them. The bill would also restrict particular uses of automated systems, including certain predictive analyses of employee behavior and inferences involving protected characteristics. In short, the proposed law would establish an important principle for workplace AI: technology may assist management, but it should not necessarily replace human judgment when an employee’s job is on the line.

A broader shift toward human oversight

SB 947 is part of a larger movement toward regulating automated employment decisions. California already treats discriminatory employment decisions made or facilitated through automated decision systems as subject to the Fair Employment and Housing Act. Other jurisdictions have adopted or proposed their own requirements involving notice, transparency, discrimination, audits, and automated employment tools. The emerging question is therefore no longer simply whether existing discrimination laws apply to AI (they do), but whether employers will face additional procedural obligations governing how AI-assisted decisions are made.

That distinction matters. An employer might ultimately reach a legally permissible employment decision, yet still face compliance issues if the process leading to that decision does not satisfy applicable notice, review, documentation, or transparency requirements. SB 947 illustrates this next phase of AI regulation. Lawmakers are increasingly focused not only on preventing discriminatory outcomes, but also on establishing guardrails around the decision-making process itself.

Practical implications for employers

Employers should consider identifying now where automated systems are being used throughout the employment lifecycle, not merely in hiring. Performance management software, productivity monitoring, attendance systems, scheduling platforms, employee analytics, and other technologies may incorporate automated decisionmaking features even when employers do not think of those products as “AI.” Employers should also understand precisely what recommendations those systems generate and how managers use them.

Human oversight should be meaningful rather than ceremonial. A manager who simply approves an AI-generated recommendation without independently reviewing the underlying circumstances may create risks under emerging laws requiring genuine human involvement. Employers should consider developing procedures identifying which employment decisions require human review, what information the reviewer must evaluate, and how that review will be documented. Vendor contracts and procurement processes should likewise address access to the information necessary to understand and, when appropriate, challenge automated recommendations.

Key takeaways: Keep humans in the loop

The legal landscape surrounding workplace AI is developing quickly, but one theme is becoming increasingly clear. Employers should be cautious about delegating consequential employment decisions entirely to technology. SB 947 has not yet become California law, and its requirements may change as it proceeds through the legislative process. Nevertheless, its approach provides a useful preview of where workplace AI regulation may be heading.

Employers using AI or automated systems should inventory those tools, determine whether they influence hiring, evaluation, discipline, termination, or other employment decisions, and establish meaningful human oversight. They should also review applicable notice and recordkeeping obligations and ensure HR professionals and managers understand that an AI-generated recommendation does not eliminate the employer’s responsibility for the resulting decision. As artificial intelligence becomes a more sophisticated workplace tool, the safest approach may be a relatively simple one: AI can assist the boss, but employers should think carefully before allowing AI to be the boss.

For more information on this topic, please contact Sunshine Fellows at sunshine.fellows@fmglaw.com or your local FMG relationship partner.

Information conveyed herein should not be construed as legal advice or represent any specific or binding policy or procedure of any organization. Information provided is for educational purposes only. These materials are written in a general format and not intended to be advice applicable to any specific circumstance. Legal opinions may vary when based on subtle factual distinctions. All rights reserved. No part of this presentation may be reproduced, published or posted without the written permission of Freeman Mathis & Gary, LLP.