Repelled at first sight? Expectations and intentions of job-seekers reading about AI selection in job advertisements

Artificial intelligence (AI) is increasingly used in personnel selection to automate decision-making. Initial evidence points to negative effects of automating these procedures on applicant experiences. However, the effect of the prospect of automated procedures on job-seekers’ pre-process perceptions (e.g., organizational attractiveness) and intentions (to apply for the advertised job) is still unclear. We conducted three experiments (Study 1 and Study 2 as within-subjects designs, Study 3 as a between-subjects design; N1=36, N2= 44, N3=172) systematically varying the information in job advertisements on the automation of different stages of the selection process (Study 1: screening stage conducted by a human vs. a non-specified agent vs. an AI; Study 2 and Study 3: human screening and human interview vs. AI screening and human interview vs. AI- screening and AI- interview). Results showed small negative effects of screening conducted by an AI vs. a human (Study 1, Study 2, Study 3), but stronger negative effects when also interviews were conducted by an AI vs. a human (Study 2, Study3) on job-seekers pre-process expectations, perceptions, and intentions. Possible reasons for these effects are discussed with special consideration of the different stages of the recruiting and selection process and explored with a qualitative approach in Study 2