Reserve Bank of India governor urges banks to make AI board-level priority

Reserve Bank of India (RBI) politician Sanjay Malhotra called connected Indian banks to attack AI arsenic a board-level strategy alternatively of basal exertion purchases and cautioned that dilatory adopters hazard losing power of the shift.

Speaking successful the opening league of FIBAC 2026, the yearly banking league organised by FICCI and the Indian Banks' Association, Malhotra said AI would beryllium the defining unit of this decade, successful the mode digitalisation marked the 2000s and liberalisation shaped the 1990s.

"The lone question present earlier america is whether you signifier the AI travel oregon you fto it signifier you by default," helium said.

According to him, AI could assistance fiscal judgement successful the aforesaid mode UPI supported payments.

Malhotra said AI was not comparable successful a elemental mode with earlier changes successful technology.

He said that portion energy accrued vigor and the net accrued connectivity, AI increases intelligence, affecting the mode banks measure and widen credit.

He said systems built connected alternate data, including currency flows, taxation filings, inferior payments and records from integer platforms, could widen entree to ceremonial recognition beyond what accepted underwriting tin achieve, particularly wherever borrowers deficiency heavy fiscal histories.

According to Malhotra, India has an borderline due to the fact that of its nationalist integer infrastructure, including Aadhaar, UPI, the relationship aggregator model and the unified lending interface.

He said AI could person a akin effect connected fiscal judgement to the 1 UPI had connected fiscal transactions, by making it immediate, elaborate and disposable astatine the past mile.

He besides named lawsuit service, vernacular dependable interfaces, operational ratio and real-time fraud detection arsenic areas wherever banks could spot speedy gains.

He said rules-based fraud systems stay down fraudsters who alteration tactics much quickly.

Malhotra said AI should enactment quality judgement alternatively than simply displace it.

He listed 7 areas of risk, including the explainability of black-box decisions, bias and exclusion carried implicit from past lending data, attraction hazard linked to a constricted radical of models and vendors, and third-party dependencies.

He besides flagged information privateness duties, cyber and adversarial vulnerabilities, and weakening of quality accountability among hazard areas.

"Fairness successful AI is not a compliance checkbox. It is simply a plan requirement," helium said, adding that last work for a bank's determination indispensable stay with the slope alternatively than a vendor oregon an algorithm.

Referring to recommendations from the RBI's committee connected the liable and ethical enablement of AI, Malhotra said banks should determination instantly connected respective fronts.