India’s central bank backs AI to widen access to loans
India’s central bank says responsible use of AI and alternative financial data could help lenders extend credit to small businesses and borrowers who struggle with traditional asse
By The Register
India’s central bank has urged lenders to explore how artificial intelligence could help provide loans to people and businesses that may struggle to qualify under traditional credit assessments.
Reserve Bank of India governor Sanjay Malhotra said AI should be responsibly harnessed by the financial sector rather than treated solely as a risk that needs to be contained.
Speaking at the FIBAC banking conference in Mumbai, Malhotra argued that new technology could help lenders assess potential borrowers using a wider range of information than conventional credit histories.
That could include cash-flow data, Goods and Services Tax filings, utility payments and other digital information.
The approach could be particularly significant for first-time borrowers, gig workers and small businesses that do not have extensive financial records or formal accounts.
Traditional lending decisions often depend on established credit histories, documented income and other financial information.
People or businesses without those records can therefore struggle to demonstrate their ability to repay even when their underlying finances may support a loan.
AI systems capable of analysing alternative sources of financial information could give lenders another way of assessing that risk.
Malhotra said the technology has the potential to help financial institutions improve efficiency, manage risk and extend access to financial services.
But the RBI is also demanding safeguards around its use.
Banks are expected to maintain an inventory of the AI systems they operate and establish governance policies approved at board level.
Financial institutions should also be able to explain automated decisions that have significant consequences for customers.
That requirement is particularly important in lending, where an opaque automated decision could affect someone’s ability to obtain credit without giving them a clear understanding of why they were rejected.
The central bank also wants AI systems to be tested before deployment and monitored after they begin operating.
Meaningful human oversight should remain in areas where an automated error could cause significant harm.
Malhotra highlighted several risks associated with greater use of AI in finance, including biased decision-making, lack of transparency, privacy concerns, cybersecurity threats and dependence on external technology providers.
Those concerns are becoming increasingly important as banks and other financial institutions introduce AI into areas previously handled predominantly by employees.
Credit assessment is potentially one of the most consequential applications.
Used effectively, AI could allow banks to consider evidence that conventional lending models overlook and potentially make finance available to viable borrowers who would otherwise be rejected.
Poorly designed systems, however, could reproduce existing biases or introduce new ones while making it more difficult for customers to challenge decisions.
The RBI’s approach therefore combines encouragement to adopt the technology with requirements for accountability, testing and human oversight.
Although the initiative is focused on India’s banking system, the underlying issue is becoming increasingly relevant internationally as financial institutions consider how far AI should be allowed to influence decisions over who receives credit.
For banks, regulators and borrowers, the question is moving beyond whether AI will be used in lending towards how those systems should be governed when their decisions have real financial consequences.