RBI Governor Sanjay Malhotra unveils AI governance roadmap, flags 7 risks

Reserve Bank of India (RBI) Governor Sanjay Malhotra has called artificial intelligence “a new way of running a bank,” while cautioning that its adoption must be tempered with robust governance. Speaking at a banking conference, Malhotra laid out a comprehensive roadmap for lenders to integrate AI responsibly, and flagged seven distinct risks that institutions must address to avoid systemic pitfalls.

AI as a Transformative Force

Malhotra emphasised that AI is not merely an incremental upgrade but a fundamental shift in how banking operations are conducted. “This is not about automating a few processes,” he said. “It is about reimagining the entire banking model, from customer service to risk assessment.” He urged banks to view AI as a strategic priority, not a back-office experiment.

The governor’s remarks come as Indian lenders increasingly deploy AI for fraud detection, credit scoring, and personalised banking. However, he stressed that speed of adoption must not outpace the ability to govern these systems.

Governance Roadmap for Banks

The roadmap outlined by Malhotra centres on accountability, transparency, and continuous monitoring. Banks must establish clear ownership of AI models, with senior management ultimately responsible for outcomes. He also called for rigorous testing of algorithms before deployment, and regular audits to ensure they perform as intended.

“If a model fails, there should be no ambiguity about who answers for it,” Malhotra said, adding that boards must receive periodic updates on AI performance metrics. He also recommended that banks maintain human oversight for high-stakes decisions, such as loan approvals or fraud alerts, to prevent over-reliance on automated outputs.

Seven AI Risks Flagged

Malhotra identified seven specific risks that banks must mitigate. These include data privacy breaches, algorithmic bias, lack of explainability, operational failures, cyber security vulnerabilities, concentration risk from third-party vendors, and potential for systemic contagion if multiple banks rely on similar AI models.

On bias, he noted that AI systems trained on historical data could perpetuate existing inequalities in credit access. “We must ensure that AI does not become a tool for exclusion,” he warned. He also highlighted the danger of “black box” models, where decisions are made without clear reasoning, making it difficult for customers to contest outcomes or for regulators to enforce compliance.

Industry Response and Next Steps

Bankers present at the event welcomed the guidance, with several noting that the RBI’s proactive stance helps set clear expectations. One senior executive said the roadmap provides a “much-needed framework” for institutions that are still experimenting with AI use cases.

The RBI has not yet issued formal regulations on AI, but Malhotra’s speech signals that a policy framework is in the works. He indicated that the central bank will consult with stakeholders before finalising any rules, balancing innovation with financial stability.

As Indian banking accelerates its digital transformation, the coming months will likely see draft guidelines from the RBI on AI governance. For now, the message is clear: AI is the future of banking, but only if it is run with discipline and foresight.