When Worlds Collide

Abstract The use of machine learning (ML) in assessments in the hiring context raises many questions regarding validity, fairness, and acceptance. This chapter looks across disciplines to provide a broad overview of key questions and the limited research addressing them. In discussing measurement and prediction, literature on construct and criterion-related validity evidence for these approaches is reviewed. Considerations in bias detection and mitigation in selection contexts are discussed. Factors related to acceptability of ML approaches to stakeholders are reviewed. The chapter concludes that while the unknowns outnumber the knowns, there are practical and psychometric questions that any practitioner can be ready to ask when considering adopting an assessment that includes an application of ML.