The Application Effectiveness of Generative AI in Credit Risk Control of Commercial Banks: A Case Study of Banks

Abstract

As commercial banks deepen digital transformation, credit risk control— a core link for asset security— increasingly adopts generative AI, though its practical application effectiveness and challenges lack systematic exploration. This study uses a case study method to analyze generative AI’s application in credit risk control at Industrial and Commercial Bank of China over the past decade, following the framework of “application scenario mapping - key indicator comparison—issue identification—suggestion formulation”. It sorts out generative AI’s core application scenarios (e.g., graphic anti-fraud, large models, intelligent agents), confirms the technology enhances risk control efficiency, reduces non-performing loan ratios, and improves model accuracy, identifies challenges like data privacy gaps, insufficient model interpretability, and high costs, and proposes suggestions from policy compliance, technical optimization, and management improvement perspectives. This research provides empirical evidence for generative AI’s value in banking risk control and offers actionable references for commercial banks balancing technological innovation and risk management.