Our study examines the intersection of two major developments in business – the enhancement of workforce diversity and the use of Artificial Intelligence (“AI”). We conduct an experiment to investigate how employee demographics moderate the effect of AI-driven performance evaluation systems (“AI-PESs”) on employees’ perceptions of system bias and employees’ behavior. We find that, relative to employee groups that are typically over-represented in business, under-represented employees perceive AI-PES to be less biased and that employees from under-represented ethnicities perceive AI-PES to be less biased than a Traditional-PES (i.e., PES reliant solely on a human manager). Further, under-represented employees are more likely to self-select into a firm and provide greater effort for a firm that uses an AI-PES vs. a Traditional-PES. Our findings suggest unintended benefits of AI use – attracting employees from under-represented groups and encouraging higher effort from these employees – collectively providing evidence that AI-PESs can influence and enhance workforce diversity.