RBI Urges Banks to Scale AI While Tightening Oversight

News

1. Temasek, Seraphim Back Pixxel in $100 Million Round

2. UN Rights Chief Warns Advanced AI Could Threaten Humanity

3. Tata-owned JLR to Cut About 4,000 Bobs Amid Tariff, China Pressure

4. Anthropic Walks Away From $6 Billion Decart Deal

5. Hyundai India Targets 20% Women in Executive Workforce by 2030

6. Security Breach Drains $320 Million in Bitcoin From Liquid Network

Reserve Bank of India (RBI) governor Sanjay Malhotra urged banks to accelerate their use of artificial intelligence, saying the technology could widen access to credit and improve productivity but would require stronger oversight of models, data and technology vendors.

Speaking at FIBAC, the annual banking conference organized by the Federation of Indian Chambers of Commerce and Industry (Ficci) and the Indian Banks’ Association (IBA), in Mumbai on Tuesday, 11 August, Malhotra said banks should invest in technology and infrastructure and train employees for an increasingly AI-driven industry rather than remain on the sidelines.

He compared AI’s potential effect on lending with the transformation brought about by the Unified Payments Interface (UPI) in digital payments.

The argument goes beyond using generative AI to automate back-office work or customer service. Malhotra said AI could change how banks assess borrowers, particularly people and businesses for whom conventional credit histories are limited or nonexistent.

Banks could use alternative data such as cash flows, Goods and Services Tax filings, utility payments and activity on digital platforms to assess new-to-credit customers, gig workers and other underserved borrowers. AI could also help lenders extend more credit to micro, small and medium-sized enterprises, improve fraud detection and offer services through Indian-language interfaces.

India already has much of the digital infrastructure required to support that shift. Malhotra pointed to Aadhaar, UPI, DigiLocker, the Open Network for Digital Commerce (ONDC), the Account Aggregator system and the Unified Lending Interface (ULI).

ULI is an RBI-led digital public infrastructure project designed to allow information from multiple data providers to flow more easily to lenders, reducing the cost and complexity of gathering information for credit decisions.

The pressure to use AI more effectively comes despite a sharp increase in banks’ technology spending.

A report released at FIBAC by Boston Consulting Group, in association with Ficci and IBA, estimated that Indian banks’ IT costs rose 6.1 times between fiscal 2015 and fiscal 2026, representing compound annual growth of about 18%.

Yet the industry’s cost-to-income ratio, a measure of how much banks spend to generate revenue, increased slightly from 47.3% in fiscal 2015 to 48.6% in fiscal 2026.

Operating expenses grew at an annualized 11.2% over the period, marginally faster than operating income at 10.9%, according to BCG’s analysis.

BCG said a decade of digitization and rising technology expenditure had therefore not fully translated into lower costs or more efficient operations. It argued that banks would have to redesign processes around AI and generative AI rather than simply add automation to existing workflows.

Malhotra, however, warned that expanding AI use creates a parallel set of risks for banks. Models can produce decisions that are difficult to explain, reproduce biases contained in historical data or expose institutions to privacy and cybersecurity failures.

Heavy dependence on a small number of model developers, cloud providers or other technology vendors could also create concentration risk across the financial system.

“The ultimate responsibility has to lie with the bank and not with a vendor or with an algorithm,” Malhotra said.

He called on banks to maintain a complete inventory of the AI models they use and put in place board-approved governance frameworks. Banks should also be able to explain important AI-driven decisions and subject their systems to red-team exercises and stress tests designed to expose weaknesses before those weaknesses affect customers or the wider financial system.

The message is consistent with a broader effort by the RBI to encourage financial-sector AI while building safeguards around it. The central bank published the report of its committee on the Framework for Responsible and Ethical Enablement of Artificial Intelligence, known as FREE-AI, in August 2025.

It is also developing a Digital Payments Intelligence Platform aimed at detecting fraud across the payments network.

The RBI’s National Strategy for Financial Inclusion 2025-30 says combating increasingly sophisticated fraud requires network-level intelligence and real-time sharing of data across payment systems, and envisages the platform becoming operational by December 2026.

State Bank of India Chairman Challa Sreenivasulu Setty, who also spoke at FIBAC, said AI could help banks take credit deeper into rural India and reach small businesses and customers whose financial histories do not fit conventional lending models.

In agriculture, he said digital records, data-driven risk assessment and satellite imagery were already creating new ways to assess borrowers and manage loan portfolios.

Setty said the challenge was to move such applications beyond limited pilots while ensuring they remained affordable and accessible. He also warned that increasingly capable technology would give fraudsters and cyber attackers more powerful tools, requiring banks to strengthen cybersecurity and retain human oversight over consequential decisions.

Indian banks are entering the AI investment cycle with unusually strong balance sheets. RBI’s June Financial Stability Report showed that scheduled commercial banks’ credit grew 14.5% in fiscal 2026, while deposits increased 11.5%. Their gross nonperforming asset ratio fell to a multi-decade low of 1.8% in March.

Banks collectively reported ₹4.05 trillion in profit after tax for the year, up from ₹3.78 trillion a year earlier. Their capital-to-risk-weighted-assets ratio rose to 17.7%, also a multi-decade high, while the core common equity Tier 1 ratio reached 15.3%.

Those buffers have become more important as banks confront risks beyond technology. Malhotra also pointed to geopolitical and trade uncertainty, although he said Indian lenders were well placed to absorb shocks.

The RBI’s latest stress tests similarly found that aggregate bank capital would remain comfortably above regulatory requirements under its adverse scenarios, even as it cautioned that global financial stability risks remained elevated.