AI-Human Labor Division Phenomenon in Credit Loan Approval

Title：AI-Human Labor Division Phenomenon in Credit Loan Approval

Speaker：Zhou Ruikai (The Chinese University of Hong Kong, Ph.D. Candidate)

Host：WANG Lin (Lingnan College, Sun Yat-sen University, Associate Professor)

Date and Time：January 3rd, 2020 (14:30-16:30)

Location：Wang Daohan Conference Room, Lingnan Hall

Language：English+Chinese

Abstract:

Will AI replace humans? Motivated by the business practice of a micro-loan company, who tries to adopt AI and gradually replace human in loan reviewing process, we model the decision process for credit loan approval, and introduce borrowers' choice over AI/human as credit reviewer. Contrasting to what managers believe, our model shows that for certain range of loans, a more accurate AI may work in favor of coexistence of AI and human, rather than using AI alone. The intuition is that the bias of human reviewer can attract those high-risk borrowers, making the borrower's choice of reviewer informative of his risk. Such “advantageous selection” can help AI make better decisions and overcome the loss from adding the inferior human reviewer. Our research sheds light on a new phenomenon of labor division when AI technologies are advancing.