Shinhan Bank "AI Now Explains How to Improve Your Credit Rating"

Integrated AI Center Research Paper Published in the Journal of the Korean Data Information Science Society

[Asia Economy Reporter Kwangho Lee] Shinhan Bank announced on the 9th that the paper titled "A Study on Credit Evaluation Models Utilizing Explainable AI Technology" by the Integrated AI Center (AICC) within the Digital Innovation Division has been approved for publication in the Journal of the Korean Data Information Science Society.

AI technology is often compared to a black box in terms of the difficulty in understanding and explaining the process of deriving results. Shinhan Bank's AICC conducted this research based on the judgment that applying AI in the financial industry requires not only high accuracy but also the ability to explain the results.

In this paper, Shinhan Bank's AICC proposed an algorithm that, along with AI-based credit evaluation algorithms, can explain the influence of each credit evaluation item on the outcome and the reasons for any fluctuations in the AI-derived results.

Through this research, when conducting AI credit evaluations, it became possible to explain to customers the impact of individual components such as salary level and debt amount on their credit rating, and to intuitively explain which factors caused an increase or decrease in their credit rating when it changes.

Additionally, Shinhan Bank's AICC collaborated with the Pattern Recognition and Machine Learning Laboratory at Korea University to jointly conduct research aimed at improving the explainability of financial AI models.

As a result, for customers who were denied loans, the AI analyzed credit evaluation items such as salary, number of delinquencies, number of utility bill delinquencies, and installment transaction ratio, and was able to suggest alternatives such as "Loans are possible if the number of delinquencies over 60 days is less than three times or if the average monthly income increases by more than 20%."

A Shinhan Bank official stated, "This research is significant in confirming that explainability can be provided when introducing AI into financial services," and added, "We will present innovative digital services centered on customers through continuous research on explainable AI (XAI)."