Responsible AI in Digital Banking: Balancing Innovation, Cybersecurity, and Ethical Governance

Abstract

Abstract. Artificial intelligence (AI) is increasingly embedded in digital banking through fraud detection, credit scoring, customer service automation, anti-money laundering monitoring, cyber threat analytics, and personalised financial services. Although these applications improve efficiency and service innovation, they also create risks related to model opacity, algorithmic bias, privacy leakage, third-party dependency, adversarial cyberattacks, inaccurate generative outputs, and unclear accountability. This article develops a systematic empirical framework for examining responsible AI implementation in digital banking by balancing AI-driven innovation capability, cybersecurity readiness, and ethical AI governance. The proposed model is grounded in the Technology-Organization-Environment framework, dynamic capability theory, socio-technical systems theory, and trustworthy AI principles. A quantitative explanatory design is recommended, using survey data from banking professionals involved in digital banking, information technology, cybersecurity, risk management, compliance, data analytics, and customer experience. The model can be tested through partial least squares structural equation modelling or moderated regression analysis. The article contributes by positioning responsible AI as an organisational governance capability rather than merely a technical compliance issue. It argues that responsible AI requires innovation routines that are supported by secure infrastructure, explainable model management, human oversight, fairness controls, privacy protection, audit trails, and board-level accountability.