India central bank head says AI can approve loans humans would have turned down, says the technology is, ‘a capability to be responsibly harnessed and not merely as a risk to be contained’ Rahimnoorali11@gmail.com (Rahim Amir) | amznusa.com

• RBI Governor Sanjay Malhotra tells banks that AI trained on alternative data such as cash flows, GST filings, and utility payments can extend credit to borrowers that manual underwriting cannot currently assess
• The pitch came bundled with caution, insisting that banks take safety and visibility measures including complete AI inventories, red-teaming before deployment, and the capacity to explain lending and fraud decisions
• The RBI insists, however, that accountability stays human; banks, not their algorithms, will be answerable to customers, auditors, and the Reserve Bank when AI-driven decisions go wrong

The governor of the Reserve Bank of India (RBI) has told the country’s banks to put artificial intelligence at the heart of their lending patterns.

The approach is based on the assumption that models trained on unconventional data can identify and leverage data points that manual underwriters would struggle to spot or justify easily.

This would, as per RBI governor Sanjay Malhotra, both strengthen the credit market and enable financial inclusion by adding a new class of borrowers previously overlooked by conventional methods.

Cautious AI optimism from one of the world’s largest central banks

The Reserve Bank of India is not a small central bank by any measure, and remarks by its governor therefore can often shape not only domestic but global markets.

His position on unlocking credit markets for users with little or no financial history by adding raw compute that considers other signals such as cash flow, GST tax filings, utility payments, and one’s digital footprint considerably changes the landscape in a part of the world where banks are traditionally more conservative than their global peers when it comes to lending.

Malhotra also noted that this is exactly the data that a first-time borrower, a gig worker, or a small enterprise without formal books lacks, even as AI enables better monitoring of debtors’ financials.

“Predictive models can identify borrowers on the cusp of default early enough to counsel rather than merely recover,” he added. “Used well, AI may be the most powerful accelerator to financial inclusion.”

It is important to point out Malhotra wanted AI to be used as an assistant here – not that AI should overrule loan officers on applications they have already rejected.

His claim is narrower and arguably more persuasive: entire categories of borrowers are effectively invisible to conventional credit assessment because the paperwork it requires does not exist, and machine learning, over other data, can make those borrowers assessable to banks and vice versa.

Malhotra also cited concerns that most modern AI models are black boxes and do not always explain their reasoning fully or adequately, and that borrowers are generally entitled to know why they were turned down. This could help keep potential algorithmic bias in check and rationalize such decisions at a time when AI models are known to be vulnerable to data poisoning attacks.

This would require a human touch, especially when it comes to accountability for such decisions, rather than the complete hands-off approach that AI enthusiasts sometimes suggest is inevitable.

To this end, the conclusion of his speech may be the most telling of what India’s central bank feels about an increased AI footprint in the banking industry: Malhotra said that the winners of the AI era will not be the fastest or heaviest adopters but the institutions that best understand what they deploy, own its outcomes, and keep customer trust, something he equated to the enduring capital of Indian banking.