The recent proceedings in Mobley v. Workday in the US District Court for the Northern District of California highlight the emerging challenges and legal risks associated with utilizing artificial intelligence in employment decision-making.
These recent advancements carry significant ramifications for the allocation of responsibility between employers and technology vendors. Employers can no longer rely on the presumption that utilizing off-the-shelf algorithmic tools shift compliance duties solely to vendors, while vendors themselves are now subject to heightened examination regarding the design and performance of their models. In this evolving landscape, conducting rigorous and systematic evaluations of AI-driven hiring systems is more critical than ever.
In the Law360 article, “Workday Case Shows Auditing AI Hiring Tools Is Crucial,” written by CRA’s Hossein Borhani, he outlines complementary empirical approaches for such evaluations: randomized internal experiments using firm-level data and internal matched-pair testing that simulates real applicants. Together, these methods provide a credible framework for assessing whether algorithmic systems operate based on legitimate job-related factors. By adopting these approaches, employers and vendors can enhance evidence-based oversight and ensure explainable algorithmic hiring.
