Stanford study finds AI hiring tools disproportionately filter out Black and Asian applicants

A Stanford-led analysis of 4 million job applications processed through the Pymetrics game-based platform across 156 employers finds clear racial disparities in hiring outcomes: Black applicants face adverse impact in about 1 in 10 roles and Asian applicants in about 1 in 20. The study also reveals many employers use identical AI models, causing the same scores across firms and creating systemic rejection across multiple applications. These findings, the largest of their kind, raise concerns about AI-driven screening and its potential to necessitate dozens of applications before receiving any human review.