Human – AI Collaboration in Credit Risk Management Enhancing Decision – Making and Risk Governance in Indian Banking Sector

Abstract

Artificial intelligence (AI) is increasingly influencing credit-risk management by enabling banks to analyse extensive borrower information, recognise risk patterns and support more informed lending decisions. Kalyani and Gupta (2023), through a systematic review of 734 studies, demonstrate the expanding role of AI and machine learning across banking activities, including credit scoring and risk management. (Springer) However, stronger predictive capability does not necessarily eliminate the need for professional judgement. Shi et al. (2022) found that machine-learning approaches can improve credit-risk prediction, while also identifying concerns relating to data imbalance, model transparency and consistency. (Springer) These concerns have increased interest in approaches that combine technological capability with human expertise. Sachan et al. (2024) show that human–AI collaboration can support augmented financial underwriting by reducing decision inconsistency and improving the quality and explainability of lending decisions. (ScienceDirect) Against this background, the present study examines human–AI collaboration in credit risk management, with particular emphasis on decision-making and risk governance in Indian banking. The study adopts a quantitative research approach and proposes primary data collection from 400 banking professionals engaged in credit assessment, lending, risk management and related functions. The conceptual framework incorporates AI predictive capability, explainability, data quality, risk-detection capability, human–AI collaboration, trust, human oversight and governance maturity as important dimensions influencing credit decision quality and risk-management effectiveness. The study argues that AI should function as an intelligent decision-support mechanism rather than an autonomous replacement for banking professionals. Effective collaboration between AI systems and human decision-makers can potentially improve analytical efficiency while preserving contextual judgement, accountability and responsible oversight.