[3Xin2] Poster session 1

[3Xin2-16] Phone Fraud Detection AI Using Factor Analysis

〇Yusuke Ohki¹, Kyohei Kato¹, Yuya Takagi¹, Hiroki Yamagishi¹ (1.Fukuoka Financial Group, Inc.)

Keywords:phone fraud detection, factor analysis, questionnaire result

In recent years, several methods have been proposed for detecting phone fraud using AI, which utilizes caller characteristics. However, with methods that rely heavily on specific characteristics of the victim (e.g. caller characteristics), detection performance deteriorates when the criminal group takes measures against those characteristics. In this study, we propose the phone fraud detection AI that is robust against countermeasures from criminal groups without assuming specific characteristics of the victim. We performed factor analysis on the results of a questionnaire about bank employees' fraud detection know-how, and the proposed method detects fraud based on that know-how. We generated hypothetical fraud victims from the questionnaire results, and the proposed method achieved better detection performance under conditions that assumed the criminal group's countermeasures.