AI Use in the Workplace: Special Mid-Year Compliance Update
Key Takeaway: Human Resource Dept. Oversight Required to Keep Pace With Rapidly Changing Technology.
This year has seen a rapid acceleration in AI integration and use amongst businesses and organizations, including Human Resources/People Operations departments.
On May 21, 2026, Governor Newson issued an executive order directing the state’s Labor and Workforce Development Agency (LWDA) to study how to mitigate AI-related effects. Newsom’s order comes alongside a robust slate of proposed state legislation aimed at regulating AI use and increasing consumer and employee protections. One of the most notable is AB 1898, which would create new notice and transparency requirements for employers around any AI use in employment decisions or employee management.
Against this backdrop, we wanted to provide five tips California employers and HR professionals can act on now as they work to ensure compliance with current obligations, and prepare for the much more substantial requirements that are likely imminent in this dynamic environment.
1. Screen AI Tools To Ensure They Are Not Being Used in a Discriminatory Manner for Employment Decisions.
As we previously reported, new California Civil Rights Department (CCRD) regulations effective last October spell out FEHA compliance obligations for employers with regard to automated decision systems (ADS) use including AI in hiring, promotion, and other employment-related decisions.
For example, employers using AI to advertise for open positions, screen job applications, or analyze interview responses cannot do so in a manner that overtly discriminates or creates a disparate impact based on a protected characteristic under FEHA. Employers must keep ADS-related records for four years, and should be aware that third party vendors can be considered their agents, thereby exposing them to greater liability.
Proposed state legislation pending this year, AB 1898 could expand these obligations to all employment decisions (not just those that fall within the ambit of discrimination protections) and employee management (including employee surveillance).
2. Conduct an Audit to Assess How AI is Being Used in Your Organization.
Another key action employers can take in the present is to audit current AI use in their organization. Initiating a broad-scale review of your organization’s AI use can help in identifying operational gaps and compliance risks, and also open the door to developing better protocols for compliant and ethical use of these technologies in your workplace. This is especially true if AB 1898 becomes law in the near future.
For example, if your organization handles any sort of confidential information, it is important to evaluate how employees are treating that data relative to any AI systems, and what protections those systems might have in place.
Employees who produce public-facing content such as social media posts could potentially run afoul of IP or defamation laws if they rely on AI-generated content without verifying its accuracy or appropriateness for use.
Also, supervisors or HR professionals who utilize AI to streamline tasks such as performance reviews may rely on systems that value certain productivity metrics but overlook key performance areas (such as communication skills), thereby creating an uneven or inaccurate assessment of an employee’s performance.
3. Draft and Regularly Revisit an Organization-Wide AI Usage Policy.
An essential component of effectively leveraging and overseeing the use of AI systems in your workplace is to develop a comprehensive, clear, and legally-compliant written AI usage policy. This can provide clarity and accountability for your workforce, and also a touchpoint for incorporating cohesive operational changes in response to shifting compliance needs.
Developing an AI usage policy can also present the opportunity to convene a staff committee, if appropriate for your organization, to assist in identifying applications and gaps with respect to how AI systems are deployed. It could also support formalizing a procedure to test any ADS technology before it is implemented for purposes of determining operational compatibility as well as conducting anti-bias testing, which can provide a defense against discrimination claims related to AI systems. Having a clear written policy can also give supervisors and HR teams a roadmap for addressing employee questions or performance issues in this context. Experienced employment counsel can assist with review or drafting of this and other HR policies.
4. Make Sure You Understand How Your Vendors Are Using AI.
The new CA CCRD regulations have already opened up a significant new area of exposure for employers relative to AI use by third party vendors, who could be considered agents of the employer relative to discrimination claims.
For example, if an employer is contracting with a third party to sort through job applications and the algorithmic screening downgrades applicants over age 40 based on proxies (such as long work histories or graduation dates), this could support a claim for disparate impact against the employer.
An employer may also rely on a third party such as a PEO who uses AI to help administer benefits or leaves, and an ADS miscalculation of something like an FMLA leave could potentially create liability for the employer.
A critical task right now for any employer working with third parties to support HR functions like recruitment or performance management is to review and ensure that they understand the terms of their vendor contracts. Employers should make sure they are aware of the systems their vendors are using, what data sources they may be drawing from, and how any data may be utilized to potentially screen, rank, or compare applicants. Employers should conduct their own assessment of the potential compliance risks associated with any individual service agreement, with the support of experienced employment counsel as needed.
5. Ensure Human Oversight of AI Systems.
The CCRD’s FEHA regulations, and likely many HR professionals’ personal experiences in using AI thus far, illuminate the importance of human oversight when utilizing ADS in the workplace. In addition to being a best practice to ensure the accuracy of outputs, human oversight of HR decisions is arguably impossible to avoid under many existing employee protections, such as those requiring an individualized assessment of a job applicant’s conviction history in relation to hiring decisions under the Fair Chance Act. Moreover, this kind of human involvement in AI usage by employers is written in as a requirement in pending legislative proposals. Given the substantial likelihood that human oversight requirements will soon be clearly mandated under future laws, employers would be best served to implement this quality control measure to support compliance now and avoid liability in the near future.
Feel free to contact us at support@clementsemploymentlaw.com with any questions about these new or potential future compliance issues, or how to update your policies or practices accordingly.
Sal Andropoulos + Wilder Clements
Senior Counsel Principal Attorney