Artificial Intelligence in Banking Fraud Detection: The Role of Accuracy, Real-Time Detection and False-Positive Management
Likitha B
Abstract
Purpose. This study examines the perceived effectiveness of artificial intelligence (AI)-based fraud detection systems in banking, focusing on real-time detection, false positives, customer experience and detection accuracy. Design/methodology/approach. The study adopts a descriptive, cross-sectional survey design using primary data from 85 respondents who regularly use digital banking services. Descriptive statistics, correlation analysis and regression analysis were used to examine relationships among the study variables. The source manuscript also reports ANOVA, although the complete ANOVA output is not provided. Findings. The reported correlation matrix indicates positive associations between overall effectiveness and accuracy (r = 0.352), real-time detection (r = 0.114), false positives (r = 0.129) and customer experience (r = 0.020). The source manuscript reports that accuracy is the only factor with a statistically significant effect on perceived effectiveness. It also reports that false positives negatively affect customer experience. Originality/value. The study provides exploratory evidence from digital-banking users on the factors shaping perceived effectiveness of AI-enabled fraud detection. It highlights accuracy as the central performance dimension while emphasising the need to balance fraud prevention with customer convenience and false-alert management.