Introduction
On March 21, 2025 [6], the California Civil Rights Council finalized regulations concerning automated decision-making systems (ADS) in employment [6]. This coincided with the introduction of a new bill, SB 7 [3] [8], known as the “No Robo Bosses Act,” proposed by State Senator Jerry McNerney. The bill and regulations aim to govern the use of AI in employment decisions, ensuring human oversight and addressing potential biases and discrimination.
Description
On March 21, 2025 [6], the California Civil Rights Council finalized regulations concerning automated decision-making systems (ADS) in employment [6], coinciding with the introduction of a new bill, SB 7 [3] [8], proposed by State Senator Jerry McNerney [3] [8], known as the “No Robo Bosses Act.” This bill aims to regulate the use of AI decision-making in various employment contexts, including hiring [1] [3] [4] [5] [8], promotions [3] [5] [6] [8], disciplinary actions [5], and terminations [3] [5] [8]. It prohibits employers from relying solely on ADS as the primary basis for significant employment decisions unless these decisions are supervised by a human. The bill broadly defines ADS as any computational process that assists or replaces human decision-making and significantly impacts individuals [8], with exemptions for basic software tools like word processors, calculators [7], spam filters [3] [8], and antivirus programs [3] [8].
The finalized regulations require employers to notify employees, contractors [3] [8], and applicants in plain language about the use of ADS at least 30 days prior to implementation [3], or by February 1, 2026 [3] [8], if already in use [3] [8]. Employers must maintain and disclose a comprehensive list of all ADS utilized in employment decisions [8]. They will also bear a greater responsibility to demonstrate that they have tested for bias and taken steps to mitigate discrimination [6], as insufficient evidence could be detrimental to their position [6]. While the regulations do not explicitly mandate bias testing of AI tools, the implications suggest that such testing may be necessary to mitigate liability and demonstrate proactive efforts to prevent unlawful discrimination [1]. Additionally, the bill aims to restrict the use of ADS that leverage personal information to predict employee behavior [5], addressing concerns about potential bias and discrimination that have led to litigation.
Workers are granted rights to access and correct data collected by ADS and can appeal decisions made with ADS involvement [3], necessitating a human review [3]. The bill includes protections against retaliation for exercising these rights [3]. Employers are required to maintain AI-related records for four years [6], which must include applications [6], personnel files [6], and data from automated decision systems [6].
As AI tools become increasingly essential for enhancing business efficiency [6], establishing an AI governance policy is crucial [1] [7]. This policy should outline a framework for responsible and ethical AI use [1], covering risk management [1] [7], bias [1] [2] [6] [7], transparency [1] [5] [7], and oversight [1] [2] [5] [7]. Organizations should implement guidelines for managing vendor relationships related to AI technology [1], including risk assessments and vendor questionnaires to evaluate potential risks before deployment [1]. Employers are encouraged to consult legal counsel to navigate potential legal liabilities and remain updated on relevant laws [6], particularly regarding the implications of the regulations on factors like accent and English proficiency, which [1] [2] [4] [6] [7] [8], while not explicitly protected under state law [1], may be associated with national origin discrimination [1].
If both the bill and the proposed regulations are enacted [3], they will complement each other in addressing the implications of ADS in the workplace [3] [8]. The final version of the regulations is expected to take effect on July 1, 2025, following approval from the Office of Administrative Law [6], while the bill is currently in the early stages [3] [8], with its first committee hearing scheduled for April 9, 2025 [3] [8]. The potential evolution of SB 7 and its influence on similar legislation in other states remain to be seen.
Additionally, the California Fair Employment and Housing Act establishes the Civil Rights Department [4], which is tasked with enforcing the act through civil actions [4]. The Department of Technology is required to inventory all high-risk ADS proposed for use by state agencies by September 1, 2024 [4]. Regulations are set forth for the development and deployment of ADS that make consequential decisions [4], mandating that developers conduct performance evaluations and provide relevant information [4], including evaluation results [4], to deployers [1] [4] [5] [6] [7]. Starting January 1, 2027 [4], deployers must disclose to individuals affected by consequential decisions made by the ADS [4], offer opt-out options [4], allow appeals of decisions [4], and submit the ADS for third-party audits [4]. The bill outlines requirements for these audits and mandates that developers [4], deployers [1] [2] [3] [4] [5] [6] [7] [8], or auditors submit unredacted performance evaluations or disparate impact assessments to the Attorney General within 30 days of a request [4], with these records exempt from the California Public Records Act [4]. Certain public entities [4], including the Attorney General [4], are authorized to initiate civil actions for noncompliance [4].
However, the bill lacks clarity regarding which HR technologies will be regulated [2], leading to significant compliance burdens for companies [2]. The broad definition of ADS may compel organizations to conduct evaluations and audits on a wide array of HR applications that utilize machine learning or data analytics [2], resulting in substantial compliance costs [2]. The requirement for performance evaluations and third-party audits [2], without clear implementation standards [2], could hinder business innovation [2], particularly for small and midsize employers that may lack the resources to manage these processes effectively [2]. The regulations would apply to any AI technology used for making consequential decisions [2], necessitating HR professionals to disclose the use of ADS to individuals affected by these decisions [2], including revealing key parameters that could disproportionately influence outcomes [2], such as scoring methodologies and performance metrics [2]. The opt-out provision could create complex processes for different worker categories [2], further complicating compliance for smaller employers [2].
While AI presents significant advantages for businesses [2], including enhanced productivity [2], the creation of high-skill job opportunities [2], improved decision-making [1] [2] [4] [5] [6] [7] [8], and tools to mitigate unintended bias [2], a balanced approach that integrates AI with human intelligence and oversight is advocated [2]. Thoughtful integration of AI technology into HR practices is essential to maximize its benefits while ensuring compliance and ethical standards are upheld.
Conclusion
The introduction of SB 7 and the finalized regulations by the California Civil Rights Council mark significant steps toward regulating AI in employment. These measures aim to ensure human oversight, mitigate bias, and protect workers’ rights. However, they also pose compliance challenges, particularly for smaller businesses [2]. As AI continues to evolve, a balanced approach that combines technological advancement with ethical considerations will be crucial in shaping the future of employment practices.
References
[1] https://www.fisherphillips.com/en/news-insights/california-regulators-adopt-new-discrimination-rules-for-automated-decision-systems.html
[2] https://www.shrm.org/topics-tools/employment-law-compliance/shrm-opposes-california-bill-restricting-ai-tools-hr
[3] https://natlawreview.com/article/no-robo-bosses-act-proposed-california-limit-use-ai-systems-employment-decisions
[4] https://calmatters.digitaldemocracy.org/bills/ca_202520260ab1018
[5] https://www.eanetpc.com/news-insights/2025/march/-no-robo-bosses-act-introduced-in-california-leg/
[6] https://www.jdsupra.com/legalnews/new-ai-laws-may-go-into-effect-as-early-4005934/
[7] https://www.jdsupra.com/legalnews/california-regulators-adopt-new-9904630/
[8] https://www.jdsupra.com/legalnews/no-robo-bosses-act-proposed-in-4118895/
