Mumbai: RBI Governor Sanjay Malhotra has told banks that artificial intelligence must not replace accountability, insisting that responsibility for decisions made with AI ultimately rests with the bank.

Speaking at the annual Fibac event in Mumbai on Tuesday, Malhotra said banks should build “meaningful human oversight” into AI systems from the design stage rather than treating it as an afterthought.

He said banks must retain the ability to explain, intervene in and override AI-generated decisions where necessary.

The governor made clear that a bank cannot blame an algorithm or technology vendor when an AI-assisted decision goes wrong. “The ultimate responsibility has to lie with the bank and not with the vendor or with the algorithm,” he said.

AI should support human judgement

Malhotra urged banks to use AI to augment rather than merely replace human judgement. He said AI could help relationship managers identify suitable products and risk indicators, allowing them to serve more customers efficiently.

He also encouraged banks to adopt AI responsibly rather than avoiding the technology because of potential risks.

The governor warned that wider use of AI could create risks involving cybersecurity, biased models and data privacy. AI systems could develop preferences or biases involving particular geographies, occupations or communities, he said.

He also highlighted the risk of “herding”, where several banks rely on a small number of AI models or technology vendors. A common bias in those systems could then spread across the financial sector.

Banks told to strengthen AI governance

Malhotra asked banks to maintain a complete inventory of the AI models being used across their operations. He also called for board-approved AI governance policies with clear accountability for outcomes.

Banks should ensure that AI systems can explain why they recommend particular decisions, while human oversight must remain in place wherever an AI error could cause significant harm to customers or affect financial stability.

Why the RBI's AI warning matters

The RBI's position comes as Indian banks increasingly use AI across areas including credit underwriting and customer service. The regulator's message is that greater automation cannot mean reduced accountability.

The emphasis on human intervention also places responsibility on banks to understand and monitor the models they deploy, particularly when they rely on external technology providers.

The RBI's approach reflects a broader regulatory challenge around AI in financial services. Banks can use AI to improve efficiency and customer service, but decisions involving credit, risk and customers can have significant consequences.

By insisting that banks remain responsible for AI-assisted decisions, the RBI is placing accountability with the regulated institution rather than allowing responsibility to shift to software developers, vendors or algorithms.