India's Central Bank Pushes AI to Expand Loan Access

Sanjay Malhotra, the governor of India’s Reserve Bank, is urging financial institutions to employ AI to approve loans that human assessors would typically reject, as The Register first reported. Speaking at the FIBAC conference in Mumbai, Malhotra stated that AI could serve as a "powerful accelerator to financial inclusion" by utilizing alternative data—such as digital footprints, GST filings, utility payments, and cash flows—to identify creditworthy borrowers who lack formal books, including small businesses, gig workers, and first-time borrowers.

According to reporting from The Register, Malhotra believes AI can also bridge accessibility gaps through voice interfaces in local languages, which is critical given that rural literacy rates are below 80 percent. Additionally, the governor noted that AI-enhanced risk models could allow regulators and banks to spot emerging financial stress more quickly than traditional financial statements allow.

However, Malhotra highlighted significant risks, specifically the "black box problem" where deep learning and generative systems cannot explain their reasoning. He warned that AI could perpetuate biases against specific communities, occupations, or geographies. To mitigate these risks, Malhotra asserted that banks must maintain human oversight and cannot blame algorithms for bad decisions, stating that "the model decided" is an unacceptable explanation for customers or auditors.

To ensure accountability, the Reserve Bank expects banks to maintain a full inventory of AI systems, establish board-approved governance policies, and conduct periodic red-team and stress-testing of systems before and after deployment.