The California Civil Rights Council has just received final approval for a comprehensive set of regulations governing the use of AI and “automated-decision systems” in employment. The regulations clarify that employers must not use AI and automated-decision systems in a manner that would discriminate against applicants or employees based on protected characteristics.
Automated-decision systems are defined in the regulations as any “computational process that makes a decision or facilitates human decision making regarding an employment benefit.” This includes any process that is “derived from and/or uses artificial intelligence, machine-learning, algorithms, statistics, and/or other data processing techniques.”
The regulations make clear that it is unlawful for an employer to use an automated-decision system or other selection criteria that discriminates against applicants or employees on a basis protected under the California Fair Employment & Housing Act or other state antidiscrimination laws (e.g., race, religion, sex, national origin). An example of discriminatory use in recruiting could be an AI system that screens out resumes with names that are perceived to belong to a particular ethnic group. The regulations will go into effect October 1, 2025.
To prepare for compliance under the new regulations, employers should consider the following:
- Assess companywide use of AI and automated-decision systems to understand their current applications.
- Conduct regular audits of AI applications/systems to identify and mitigate any biases or other unintended consequences.
- Examine the effects of current AI applications/functions specific to human resources and recruiting practices. For example, is AI being used to assist in resume review, creating interview questions, or managing performance evaluations. If so, how are these processes being impacted?
- Ensure the company has created and distributed a comprehensive AI Use Policy to all employees.
- Train all employees on the ethical use of AI and the company’s AI Use Policy.
- Establish a review process for AI decisions affecting employment outcomes.