Adoption of Artificial Intelligence in the Banking Sector: A Systematic Review and Thematic Synthesis

Abstract

Banks are increasingly adopting artificial intelligence (AI) and machine learning (ML) to automate routine tasks, assess creditworthiness, manage risks, and support strategic decision-making. This study conducts a systematic literature review of 91 Scopus-indexed publications following PRISMA 2020 guidelines, with study quality assessed using MMAT 2018 and CASP frameworks. The review examines theoretical perspectives on AI adoption by banks and consumers, factors influencing implementation, and implications for financial performance, credit risk management, and regulatory compliance. The findings indicate that institutional AI adoption research commonly uses institutional theory, Technology–Organization–Environment (TOE), Resource-Based View (RBV), and Resource Orchestration Theory (ROT). In contrast, consumer adoption studies frequently apply the Technology Acceptance Model (TAM), UTAUT2, and Diffusion of Innovation theory. Successful AI implementation depends on technological infrastructure, employee capabilities, fintech competition, and regulatory certainty. AI adoption can improve financial performance, corporate governance, compliance efficiency, and creditworthiness assessments. However, challenges involving data security and privacy, algorithmic transparency, and financial system stability remain significant concerns. This review provides insights for banking executives and policymakers to develop effective, transparent, and ethically responsible AI adoption strategies.