Voice AI in Firms: A Natural Field Experiment on Automated Job Interviews | Brian Jabarian and Luca Henkel

Talent Acquisition
Tags:

If you find value in content like this, sign up for my Talent Edge Weekly newsletter. 

This new 89-page research paper examined what happens when AI voice agents replace human recruiters in job interviews. Partnering with a recruitment firm, the researchers conducted a large-scale field experiment with 70,000 applicants who were randomly assigned to be interviewed by human recruiters, AI voice agents, or given the latitude to choose between the two. Regardless of the interview type, questions were consistent, and final hiring decisions always rested with human recruiters who reviewed transcripts, recordings, and standardized test scores. Contrary to recruiter forecasts that AI would lower interview quality and harm outcomes, AI-led interviews improved results, such as 12% more job offers and 18% more job starts. AI also elicited more relevant information (6.8 topics covered versus 5.5 in human-led interviews), leading recruiters to score AI-interviewed applicants higher. Interestingly, when given the choice, 78% of applicants opted for an AI interview over a human one, often citing convenience and reduced anxiety. However, those who selected AI interviews generally scored lower on standardized test scores. Another point is that 5% of applicants discontinued an AI interview in progress, and 7% encountered technical glitches that added to their stress. As with all studies, this research has limitations; however, it provides a useful data point to consider alongside broader research. I also found this 17-minute AI-generated podcast summarizing the findings, implications, and limitations of this study.