Artificial Intelligence (AI) in bank lending: moving beyond heuristic-based credit

Artificial Intelligence (AI) is transforming loan approval practices in the global banking sector. Its introduction has reshaped the lending landscape by redefining decision-making processes, positioning AI as a decision-support mechanism in which routine assessments are automated. Human expertise is directed towards exception handling, ethical oversight, and strategic risk assessment. In Malaysia, regulatory emphasis on responsible AI adoption highlights the need to balance efficiency gains with transparency, explainability, and accountability (Bank Negara Malaysia, 2025). Effective governance frameworks seek to ensure that AI reduces behavioural bias without introducing systematic risks or undermining trust in the banking system. Loan approval has traditionally been a judgement- intensive process, due to its heavily reliance on human discretion. Credit officers typically evaluate borrowers using financial ratios, credit history reports from various agencies, and qualitative impressions formed through documentation reviews and interviews. While professional judgement plays an important role, behavioural finance research has long established that human decision-making is subject to heuristic, cognitive limitations and emotional biases, particularly under conditions of uncertainty and time pressure (Kahneman, 2011). The emergence of AI in banking reflects an effort to reduce these decision-making risks by introducing data-driven, consistent, and scalable lending systems. Both globally and in Malaysia, increasingly deploying AI to support or automate loan approval decisions, with the aim of improving efficiency, accuracy, and governance (Bank Negara Malaysia, 2025).