A Secondary Study on Hybrid Human–AI Models for Enhancing Credit Risk Assessment in the Indian Banking Sector: Evidence from SBI

Abstract

Credit risk assessment is a fundamental function in the banking sector, directly influencing financial stability and loan performance. Traditional credit evaluation methods, primarily based on financial ratios, credit history, and human judgement, often face limitations in handling large-scale and complex datasets. With the rapid advancement of artificial intelligence (AI) and machine learning (ML), financial institutions are increasingly adopting data-driven approaches to improve the accuracy and efficiency of credit risk prediction.
This secondary research paper explores the effectiveness of hybrid Human–AI models in enhancing credit risk assessment, with a specific focus on the State Bank of India (SBI). The study is based on an extensive review of existing literature, including academic journals, RBI reports, and industry publications. It examines how AI-based models such as logistic regression, random forest, and neural networks contribute to improved predictive accuracy while identifying their limitations, including algorithmic bias and lack of transparency.
The research highlights that integrating human expertise with AI systems creates a balanced decision-making framework. While AI provides data-driven insights and predictive analytics, human judgement ensures contextual understanding, fairness, and accountability. This hybrid approach not only improves risk prediction but also enhances operational efficiency and
reduces non-performing assets (NPAs).
The findings suggest that hybrid human–AI models can significantly strengthen credit risk management in large public sector banks. The study also emphasises the need for regulatory guidelines and strategic implementation to ensure responsible and effective use of AI in banking.