AI Capability, Risk Management and Financial Performance: Evidence from Islamic Banks

Abstract

This study investigates the relationship between artificial intelligence, AI capability and financial performance in Islamic banks and examines whether risk management serves as a mediating mechanism in this relationship. Drawing on the resource-based view and dynamic capabilities theory, the study further examines whether Shariah board expertise and fintech regulatory intensity strengthen the effectiveness of AI capability in improving risk outcomes. The study employs a ten-year unbalanced panel of 148 Islamic banks operating across 16 OIC and non-OIC countries during 2014–2023. AI capability is operationalized using three complementary indicators: AI-related patents, AI-related job postings, and AI-related textual intensity in annual reports. Financial performance is measured using return on assets (ROA), return on equity (ROE), and Tobin's Q, while risk management is captured through the Z-score, non-performing financing (NPF), and operational loss events. The empirical analysis employs bank and year fixed-effects models, while two-stage least squares (2SLS) and system generalized method of moments (system-GMM) are used to address potential endogeneity. Bootstrap procedures are applied to test the mediating effects. The results indicate that AI capability is positively and significantly associated with ROA, ROE, and Tobin's Q. AI capability is also associated with improved risk outcomes, reflected in higher financial stability and lower NPF and operational losses. The inclusion of risk-management variables reduces the direct effect of AI capability on financial performance while leaving it statistically significant, indicating partial mediation. Bootstrap estimates further confirm significant indirect effects through risk management. The interaction analyses show that the beneficial effect of AI capability on risk management is stronger in Islamic banks with greater Shariah board expertise and in countries characterized by stronger fintech regulatory environments. The findings remain consistent across alternative AI measures and estimation techniques. Islamic banks can enhance the economic value of AI by integrating AI-enabled financing assessment, fraud detection, operational monitoring, and compliance technologies into their risk-management systems. Regulators should develop AI governance frameworks that address algorithmic transparency, data governance, cybersecurity, model risk, and Shariah compliance. This study contributes to the emerging literature on AI and Islamic finance by developing an integrated AI capability–risk management–financial performance framework and providing multi-country longitudinal evidence on the mechanisms and institutional conditions through which AI creates value in Islamic banking. The findings also connect AI-enabled risk management with the broader objective of wealth preservation (Hifz al-Mal) within the Maqasid al-Shariah framework.