California Public Employees AI Notice Bill
Public employees: notice: artificial intelligence performing service within scope of work
United States • California
RAI-US-CA-AB26560-2026AB 2656
AB 2656 requires California public employers to give 45 days' notice to employee unions before using generative AI for tasks within employees' scope of work.
Overview
California Assembly Bill 2656 (AB 2656) is a pivotal legislative proposal aimed at establishing robust transparency and collaboration requirements for the deployment of artificial intelligence (AI) within the state's public sector workforce. Introduced to address the rapid integration of generative artificial intelligence (GenAI) into government operations, the bill mandates that state and local public employers provide a minimum of 45 days' written notice to recognized employee organizations (labor unions) before developing, purchasing, implementing, or utilizing GenAI systems that perform services within the scope of work of represented job classifications. This proactive measure is designed to ensure that public employees and their representatives are not only informed but also actively engaged in the decision-making process concerning technological advancements that could significantly alter their roles and working conditions. The bill's core intent is to foster a collaborative environment where the benefits of AI can be harnessed responsibly, mitigating potential negative impacts on the workforce, such as job displacement, skill obsolescence, or changes in work processes without prior consultation.
The urgency for AB 2656 stems from the accelerating pace of AI adoption across various governmental functions, from administrative tasks to specialized services like law enforcement. Organizations like the Peace Officers Research Association of California (PORAC), a key sponsor, have emphasized the critical need for such legislation to prevent public employees from being 'left in the dark' as AI technologies become increasingly prevalent. By institutionalizing a clear notice requirement, AB 2656 seeks to embed accountability and thoughtful planning into the process of AI integration. This framework is intended to facilitate meaningful dialogue, allowing unions to raise concerns, propose safeguards, and negotiate terms related to wages, hours, and working conditions that may be affected by AI deployment. Ultimately, the bill aims to ensure that the introduction of AI in the public sector is a managed transition that respects employee rights and leverages technology to enhance public service delivery without compromising the welfare of the public workforce.
Definitions
AB 2656 carefully leverages existing California Government Code to define its core technological terms, ensuring consistency and avoiding ambiguity within the state's legal framework. The bill specifically cross-references Government Code Section 11546.45.5 for its definition of "artificial intelligence." Under this existing statute, "artificial intelligence" is broadly understood as an engineered or machine-based system that demonstrates varying degrees of autonomy. Such systems are characterized by their ability to infer from the input data they receive and subsequently generate outputs that can exert influence over physical or virtual environments. This comprehensive definition encompasses a wide spectrum of AI capabilities, from basic automation to complex decision-making systems.
The bill further refines its scope by defining "generative artificial intelligence" (GenAI) as a specialized subset of AI systems. GenAI is explicitly described as an artificial intelligence system engineered to produce derived synthetic content. This synthetic content is diverse, including but not limited to text, images, video, and audio, all of which are designed to mimic the structural and characteristic patterns found in the system's training data. This distinction is crucial because GenAI's capacity to create novel content has profound implications for tasks traditionally performed by human employees, making it a primary focus for workforce impact legislation. While these AI definitions are robust, a critical aspect of AB 2656's application hinges on the term "scope of work." The bill itself does not provide a specific definition for "scope of work," leaving its interpretation to existing labor law precedents, collective bargaining agreements, and potentially future judicial or administrative rulings. This lack of explicit definition could lead to initial challenges in determining precisely when the notice requirement is triggered, necessitating careful consideration during implementation and potential future legislative or regulatory clarification.
Governance and Institutional Framework
The governance and institutional framework underpinning AB 2656 is primarily procedural, designed to integrate AI deployment considerations into California's established public sector labor relations system rather than creating new regulatory bodies. The central responsibility for compliance rests with "state and local public employers" across California. These entities are mandated to proactively identify situations where generative artificial intelligence (GenAI) will be developed, purchased, implemented, or utilized to perform services that fall within the existing scope of work of their represented employees. This places a significant onus on public agencies to conduct internal assessments and maintain awareness of their technological acquisitions and deployments.
The core governance mechanism is the strict requirement for these employers to provide timely written notice to the "recognized employee organization" that represents the affected job classification. These recognized employee organizations, typically public sector labor unions, are thus empowered as key stakeholders in the AI integration process. Their role is to receive the mandatory notice, review the proposed AI deployment, and engage in discussions or negotiations with the employer regarding its potential impacts on their members' wages, hours, and working conditions. While AB 2656 does not establish a new oversight agency, the enforcement of its provisions would naturally fall under the existing jurisdiction of the Public Employment Relations Board (PERB). PERB is the state agency responsible for administering California's collective bargaining statutes for public employers and employees. Any disputes arising from an employer's failure to provide the mandated notice, or from subsequent disagreements over the impacts of AI deployment, would likely be adjudicated through PERB's established unfair labor practice procedures, leveraging existing institutional structures to ensure compliance and resolve conflicts.
Key Focus Areas
AB 2656 strategically targets several critical dimensions at the intersection of artificial intelligence and public employment, aiming to create a balanced and equitable transition as AI technologies become more pervasive. Its foremost emphasis is on transparency and advance notification. The bill explicitly mandates that public employers furnish a minimum of 45 days' written notice to recognized employee organizations before undertaking any action involving generative AI (GenAI) that falls within the scope of work of represented employees. This substantial notice period is not merely a formality; it is meticulously designed to provide employee representatives with sufficient time to thoroughly understand the specifics of the proposed AI deployment, assess its multifaceted potential impacts on their members, and prepare for meaningful engagement and dialogue with employers. This proactive disclosure is intended to prevent unilateral decisions that could disadvantage the workforce.
A second crucial focus is the specific scope of AI application, with a deliberate concentration on generative AI systems. The legislation acknowledges the uniquely transformative capabilities of GenAI, which can autonomously produce synthetic content such as text, images, video, and audio. This generative capacity directly impacts tasks traditionally performed by human workers, raising concerns about job redefinition, skill shifts, and potential displacement. By specifically targeting GenAI, the bill addresses the most immediate and tangible shifts in work processes that these advanced AI systems are poised to introduce. Furthermore, the bill strongly emphasizes employee engagement and collaboration. The mandated notice requirement is fundamentally designed to facilitate robust discussions and negotiations on critical aspects such as wages, hours, and working conditions. This ensures that the introduction of AI is not an imposed change but rather a collaborative process that respects employee rights, maintains job security where possible, and explores opportunities for upskilling or reskilling. This proactive engagement is deemed essential for fostering a smooth, fair, and equitable transition as AI increasingly integrates into government operations, ensuring that technological progress benefits both public service delivery and the public workforce.
Implementation Framework
The implementation framework for AB 2656 is structured around a clear, mandatory notification protocol, designed to be integrated seamlessly into existing public sector labor relations practices. The process is triggered when a state or local public employer forms an intent to develop, purchase, implement, or utilize any generative artificial intelligence (GenAI) system. A critical condition for this trigger is that the GenAI system must be intended to perform services that fall within the established scope of work of a job classification represented by a recognized employee organization. This necessitates an internal assessment by the employer to determine the AI system's functional capabilities and its overlap with existing human roles.
Once this condition is met, the employer is legally obligated to provide a comprehensive written notice to the relevant recognized employee organization. This written notice must be delivered no less than 45 days before the employer proceeds with any of the aforementioned actions related to the GenAI system. The 45-day period is a crucial window, providing the recognized employee organization with ample time to: 1) review the details of the proposed AI deployment, including its intended functions, potential impacts on job duties, and any associated changes in work processes; 2) understand its implications for their members, including potential changes to workload, required skills, or job security; and 3) prepare for and engage in substantive discussions or negotiations with the public employer. While the bill mandates the notice, it implicitly relies on California's existing labor-management relations processes, including collective bargaining obligations, to facilitate the subsequent dialogue and address any concerns raised by the employee organization. This approach ensures that AI deployment is handled transparently and collaboratively, leveraging established legal frameworks for addressing changes in working conditions rather than creating an entirely new administrative burden.
Monitoring and Evaluation
AB 2656's primary legislative focus is on establishing a crucial pre-deployment notice requirement, rather than detailing explicit mechanisms for the ongoing monitoring and evaluation of AI systems once they are implemented. The bill's core intent is to ensure transparency and robust employee organization involvement at the foundational stages of AI adoption within California's public sector workplaces. Consequently, the bill does not prominently outline specific provisions for state-level monitoring of AI system performance, mandates for post-implementation impact assessments, or requirements for regular audits of AI usage by a newly created, dedicated oversight body. This design choice reflects an approach to integrate AI governance into existing labor relations structures.
However, the spirit and practical implications of the bill suggest that any subsequent monitoring and evaluation would largely occur within the established frameworks of public sector labor relations. Recognized employee organizations, having received the initial 45-day notice, are strategically positioned to monitor the actual implementation and observe the practical impact of GenAI on their members' work. This ongoing observation by employee representatives serves as a decentralized, yet effective, form of oversight. Should any adverse effects emerge, such as unintended job redefinition, increased workload, or deviations from terms discussed during the notice period, these could potentially trigger formal grievances, unfair labor practice charges filed with the Public Employment Relations Board (PERB), or necessitate further negotiations under existing collective bargaining laws. While the bill itself does not impose new reporting requirements to a central state authority regarding AI system performance or societal impact, the continuous dialogue and potential for formal dispute resolution between public employers and employee organizations, initiated by the mandatory notice, would serve as an implicit and dynamic mechanism for evaluating the practical implications of AI deployment on the workforce and ensuring accountability.
Penalties, Liability, and Appeals
As currently drafted and analyzed, AB 2656 primarily focuses on establishing a mandatory procedural obligation for public employers—the 45-day written notice requirement—and does not explicitly introduce new, specific penalties, liability provisions, or a formal appeals process directly within its text for non-compliance. The bill's intent is to embed this notice requirement within the existing framework of California's public employment labor laws. Therefore, any recourse available to a recognized employee organization in the event of an employer's failure to provide the mandated notice, or other related violations, would fall under the comprehensive statutes and regulations governing public employment relations in California.
In scenarios where an employer fails to comply with AB 2656's notice requirement, a recognized employee organization would typically pursue remedies through established labor relations mechanisms. The primary avenue for such enforcement would be the Public Employment Relations Board (PERB). PERB is the state administrative agency charged with administering and enforcing the collective bargaining statutes that cover California's public employers and employees. A recognized employee organization could file an unfair labor practice charge with PERB, alleging that the employer has violated its duty to bargain in good faith or its obligation to provide notice regarding changes to terms and conditions of employment, which AB 2656 would now explicitly include for GenAI deployment. Violations of mandatory notice or bargaining obligations under existing law can lead to a range of remedies imposed by PERB, including, but not limited to, cease and desist orders, orders to bargain in good faith, reinstatement of employees, back pay, or other actions deemed necessary to make the aggrieved party whole and restore the status quo. While AB 2656 does not create novel penalties, its requirements become an integral part of an employer's existing obligations under California's robust public employment relations statutes, thereby making non-compliance subject to the significant enforcement powers of PERB and the broader legal system.
Relationship to Other Instruments
AB 2656 is not an isolated piece of legislation but rather a carefully crafted instrument that builds upon and interacts with several other significant legislative and executive directives in California concerning artificial intelligence and public employment. A notable precedent for AB 2656 is Assembly Bill 339, which was enacted into law and became effective on January 1, 2026. AB 339 established a similar 45-day notice requirement, obligating public agencies to inform labor unions in advance of soliciting contracts for services that fall within the scope of work traditionally performed by represented employees. This prior legislation clearly signals a consistent legislative intent to ensure advance notification and union engagement in matters affecting public sector work, providing a strong foundation for AB 2656's specific focus on AI.
Furthermore, AB 2656 operates within the broader context of Executive Order No. N-12-23, issued by the Governor. This executive order mandates that specified state agencies collaborate closely with their workforce to develop a comprehensive report. This report is tasked with examining both the beneficial use cases and the potential risks associated with the deployment of generative AI by the state. Crucially, the executive order also requires regular assessment and updates to this report, explicitly stipulating consultation with state government workforce organizations. Complementing this, Chapter 928 of the Statutes of 2024, known as the Generative Artificial Intelligence Accountability Act, further reinforces these requirements. This act mandates that state agencies update the aforementioned report as needed in response to significant technological developments and, critically, to consult with organizations representing state exclusive employee representatives. The act also encourages state agencies to explore procurement and enterprise use opportunities for GenAI to enhance efficiency and equity in public services. Therefore, AB 2656 aligns seamlessly with this overarching state strategy, which aims to integrate AI cautiously and responsibly while prioritizing transparency, safeguarding public employees' interests, and embedding robust notice requirements directly into the employment relations framework, ensuring a holistic approach to AI governance.
National/Federal Alignment
California AB 2656 represents a pioneering state-level legislative effort to specifically address the workforce implications of artificial intelligence, particularly generative AI, within the public sector. This initiative positions California at the forefront of developing granular regulatory frameworks for AI's impact on employment, an area where comprehensive federal alignment is still largely in its formative stages. While the federal government has indeed issued significant executive orders and strategic plans concerning AI development and use, such as the Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence (October 2023), these federal directives generally provide high-level principles, broad guidelines, and strategic objectives rather than prescribing specific notice requirements for public sector employment at the state or local level. Federal guidance typically focuses on overarching themes like AI safety, security, privacy, equity, and innovation across various sectors, including government, but it often defers the intricate specifics of labor relations, workforce impact, and collective bargaining to individual federal agencies or, more commonly, to state and local jurisdictions.
Consequently, AB 2656 can be viewed as a significant contribution to the broader national discourse on AI governance, offering a concrete and actionable model for how states can proactively address the immediate and tangible concerns of their public workforces. It demonstrates a legislative approach that effectively integrates AI deployment considerations within existing labor relations frameworks, a strategy that could potentially inform future discussions at the federal level or serve as a blueprint for other states grappling with similar challenges. While there is currently no direct federal mandate for a 45-day notice to public employee unions regarding GenAI deployment, AB 2656's core principles of transparency, employee engagement, and responsible AI development resonate strongly with broader federal interests in ethical AI use and the mitigation of potential negative impacts on employment. These themes are consistently present in federal AI policy discussions, indicating a philosophical alignment even in the absence of direct regulatory parallelism. Thus, California's bill exemplifies a progressive state-led effort to operationalize responsible AI principles in a critical sector, potentially influencing the evolution of national AI policy.
Implementation Timeline
| Milestone | Date | Notes |
|---|---|---|
| Introduced | 2026-02-20 | Introduced by Assembly Member Petrie-Norris. |
| First Committee Hearing (Assembly P. & C.P.) | 2026-04-16 | From committee: Do pass and re-refer to Com. on P. E. & R. (Ayes 14. Noes 1.). |
| Second Committee Hearing (Assembly APPR) | 2026-04-22 | From committee: Do pass and re-refer to Com. on APPR. (Ayes 7. Noes 0.). |
| Passed Assembly | 2026-05-26 | Passed Assembly 72-2. |
| Referred to Senate Committees | 2026-05-26 | Ordered to the Senate, referred to Coms. on L., P.E. & R. and P., D.T., & C.P. |
| Approved by Senate Privacy Committee | 2026-06-17 | Approved by Senate Privacy (5-0). |
| Approved and Re-referred to Senate Appropriations | 2026-07-01 | Approved and re-referred to Senate Appropriations. |
| Current Status | 2026-07-06 | Under review in Senate Appropriations Committee. |
Compliance Checklist
| Check | Required Action |
|---|---|
| Identify Generative AI Use | Determine if the public employer plans to develop, purchase, implement, or utilize any generative AI (GenAI) system. |
| Assess Scope of Work Impact | Verify if the GenAI system will perform a service that is within the scope of work of a job classification represented by a recognized employee organization. |
| Prepare Written Notice | Draft a comprehensive written notice detailing the intent to deploy GenAI, including its intended functions and potential impacts. |
| Provide 45-Day Advance Notice | Deliver the written notice to the recognized employee organization(s) no less than 45 days before any action (development, purchase, implementation, utilization) related to the GenAI system. |
| Engage with Employee Organizations | Be prepared to engage in good-faith discussions or negotiations with the recognized employee organization(s) regarding the GenAI deployment, its impacts on wages, hours, and working conditions, and potential mitigation strategies. |
Sources and References
| Source | Type |
|---|---|
| AB 2656 - Assembly Bill Policy Committee Analysis - CA.gov | government |
| California Assembly Bill 2656 - LegiScan (Status) | legal |
| Bill Text: CA AB2656 | 2025-2026 | Regular Session - LegiScan | legal |
Related Regulations
California Assembly Bill 2545: Report: Labor Force Impact: Artificial Intelligence
California, United States86% similar
Health care services: artificial intelligence.
California, United States86% similar
Health care professions: deceptive terms or letters: artificial intelligence.
California, United States83% similar
California AB 2655 - Defending Democracy from Deepfake Deception Act of 2024
United States83% similar
California AI Transparency Act
California, United States83% similar
© Regulations.AI — created on 06-Jul-2026 using Gemini 2.5 Flash