AB 1898 would add a new part to the Labor Code creating “Artificial Intelligence Transparency at Work.” The bill requires employers to give written notice when a workplace AI tool is used to assist with employment-related decisions or to surveil workers. The notice must be provided before first deployment, to new hires, and for existing tools by a set deadline, and it must be written in plain language and in the language used for routine worker communications. The notice must describe, among other things, the purpose of the tool, the kinds of decisions it may affect, what worker data it collects, how it is stored and used, who can access it, what surveillance methods are involved, the model creator, quotas imposed by the tool, whether jobs may be automated or replaced, training provided, and a summary of any risk assessments.
The bill also requires employers to maintain an updated annual list of all workplace AI tools in use and to disclose whether newly added tools may replace or automate jobs or job tasks and on what timeline. It defines key terms broadly, including “employer,” “worker,” “worker data,” “automated decision system,” and “workplace surveillance tool,” and it expressly includes public-sector employers such as state agencies, local governments, schools, and universities. The bill authorizes enforcement by the Labor Commissioner, a public prosecutor, or an affected worker or union representative through civil action, and it provides remedies including injunctive relief, attorneys’ fees, punitive damages, and penalties of up to $500 per employee per violation.
If enacted, AB 1898 would create a new transparency and notice regime in California employment law for AI-driven hiring, discipline, scheduling, productivity, and surveillance tools. It would add new employer compliance obligations, expand potential liability for noncompliance, and apply statewide, including to charter cities, while preserving local ordinances that provide equal or greater worker protections. The bill also references compliance with other state law, including the California Consumer Privacy Act, signaling overlap with existing privacy and data-handling requirements.
The general sentiment reflected in the available voting history appears favorable but not unanimous. The bill received a unanimous 7-0 do pass vote in one committee and later a 10-3 do pass vote in another committee, suggesting substantial support for worker transparency around AI use, but also some opposition or concern. The bill was later held under submission in committee, indicating that despite positive votes, it did not advance cleanly at that stage.
The main points of contention likely center on the breadth and burden of the disclosure requirements. Employers may view the bill as imposing detailed notice, recordkeeping, and timing obligations, including a requirement that workers sign notices before the tool may be used. Labor and worker advocates are likely to support the bill’s emphasis on transparency, notice, and accountability for AI-based surveillance and decisionmaking, while critics may be concerned about administrative complexity, litigation exposure, and the bill’s broad application to both private and public employers.
AB 1898 would add new sections to Division 2 of the Labor Code establishing a statewide workplace AI transparency framework. It would require employer disclosures about AI-assisted employment decisions and workplace surveillance, annual inventories of AI tools, and detailed information about data collection, model use, quotas, automation impacts, and risk assessments. The bill would also create enforcement authority for the Labor Commissioner and public prosecutors, and a private right of action for workers and their exclusive representatives, with penalties and damages available for violations. It applies broadly to private employers and public entities, including charter cities, while not preempting stronger local protections.
The available vote history suggests the bill was generally well received in committee, with strong support reflected in a 7-0 vote and a later 10-3 vote to advance it. At the same time, the fact that it was held under submission indicates unresolved concerns or procedural hesitation. Overall, the sentiment appears supportive of the bill’s worker-protection and transparency goals, but not without some resistance to its compliance and enforcement structure.
The likely areas of contention are the scope and operational burden of the bill’s requirements. Employers may object to the breadth of the definition of workplace AI tools, the detailed content required in notices, the annual list of tools, and the requirement that workers sign acknowledgments before use of covered tools. There may also be concern about the bill’s private enforcement provisions, penalties, and potential exposure to litigation. Supporters, likely including labor advocates and privacy-focused stakeholders, would emphasize the need for advance notice, transparency about surveillance and automated decisionmaking, and worker protection from hidden AI-driven employment decisions.