A study has found that artificial intelligence (AI) interviewers demonstrate superior job performance compared to human interviewers. The existing premise that humans would outperform AI in conversation-centered interviews appears to be shaken.
On the 28th (local time), Bloomberg reported that a collaborative research team from the University of Chicago Booth School of Business and Erasmus University in the Netherlands found that candidates selected by AI interviewers had higher success rates and retention rates compared to those chosen by human interviewers.
The study was conducted in collaboration with a recruitment agency, targeting a total of 67,000 job seekers applying for entry-level customer service positions at corporations in the Philippines. Job seekers participated in interviews by choosing their preferred interviewer from either an AI interviewer or a human interviewer, or being randomly assigned, and selections were made based on numerical evaluations provided by each interviewer.
In particular, the AI interviewer reported an average 12% higher success rate than the human interviewer, and the retention rate for successful candidates was also about 17% higher. This is interpreted as a result of the AI interviewer being able to cover a wider range of key topics within a limited time while not experiencing fatigue, thus maximizing efficiency.
Job seekers also showed a satisfactory response. According to the researchers, candidates who were given the choice of interviewer preferred the AI interviewer over the human interviewer four out of five times, and among those who interviewed with the AI interviewer, 70% rated their interview experience as "positive." This is double the rate of the same response given to human interviewers.
However, there were also those who felt uncomfortable with AI interviews. About 5% of job seekers abandoned the interview midway due to discomfort in conversing with the interviewer, and 7% reported having difficulties during the interview process due to technological issues. Additionally, numerous feedbacks mentioned that the interviewer's tone was not natural.
The researchers emphasized that the cost-effectiveness and job performance of implementing AI interviews could vary depending on the scale of recruitment and the salaries of hiring managers. For example, in small businesses where hiring managers' salaries are set low, even if AI interviews are implemented, recovery of the investment costs could be challenging, as AI interviewers typically take twice as long for result reviews compared to humans, which could actually reduce efficiency.
On the other hand, in large companies with many applicants or higher salaries for hiring managers, AI interviews could be an efficient alternative as they allow for flexible scheduling and quick progression through stages. Additionally, for customer-facing roles with high turnover rates, AI interviews could maximize the retention rates and job performance of successful candidates, lowering the costs of hiring new employees.
Brian Zabarian, a researcher at the University of Chicago, forecasted that a hybrid model where AI handles initial screening and interview processes while humans make final decisions is likely to spread in future hiring processes. He noted that, "AI implementation should be carefully assessed based on situational scenarios and operational effects," adding that "continuous validation is necessary to realize investment benefits beyond just AI introduction."