The Governor’s Call for Responsible Innovation
In a significant address at the recent FIBAC 2026 conference in Mumbai, RBI Governor Sanjay Malhotra laid out the central bank's vision for the future of banking. He urged lenders to harness the transformative power of Artificial Intelligence, comparing
its potential impact on this decade to that of digitalisation in the 2010s. However, his endorsement came with a strong and clear warning. The Governor stressed that this technological leap must be anchored in strong governance, accountability, and meaningful human oversight. He was unambiguous, stating that the RBI views AI as a capability to be responsibly harnessed, not merely a risk to be contained. This dual message encourages innovation while firmly placing the responsibility for its consequences on the banks themselves, making it clear that blaming an algorithm for a decision will not be an acceptable explanation for customers or regulators.
The Double-Edged Sword of AI
Artificial Intelligence is already changing the face of Indian banking. From AI-powered chatbots like HDFC's EVA handling customer queries to sophisticated algorithms detecting fraudulent transactions in real-time, the technology is woven into daily operations. The promise is immense. AI can dramatically lower the cost of underwriting loans, making it feasible to extend credit to gig workers and small businesses who lack formal financial histories. By analysing alternative data like cash flow and utility payments, AI could significantly boost financial inclusion. Furthermore, AI-powered systems operating in various Indian languages can make banking more accessible to a wider population. However, the very power that makes AI so promising also makes it risky. If deployed carelessly, the Governor warned, it could create new forms of financial exclusion and introduce instability into the system at a pace that is difficult to manage.
Decoding the Key Risks
The RBI Governor highlighted several critical risks that require robust guardrails. First is the 'black box' problem, where complex AI models make decisions—like rejecting a loan—without being able to provide a clear reason. Another major concern is algorithmic bias. An AI trained on historical data could inadvertently learn and amplify existing societal biases against specific communities or regions, leading to discriminatory outcomes. There is also the 'herding' risk: if most banks rely on a few dominant AI models from the same vendors, a single flaw or bias could trigger a system-wide crisis. Beyond these, data privacy and new cybersecurity vulnerabilities remain paramount concerns. The Governor emphasised that ultimate accountability must remain with humans, ensuring AI is used to augment, not simply replace, human judgment and oversight.
What 'Guardrails' Could Look Like
In response to these risks, the RBI is signalling a more hands-on regulatory approach. Governor Malhotra called for banks to establish board-approved AI governance policies that outline clear lines of accountability. He also stated that banks should maintain a complete inventory of all AI models in use, a move designed to increase transparency and regulatory oversight. Future regulations may likely focus on model explainability, requiring banks to be able to interpret and justify AI-driven decisions. Stricter rules around data usage, third-party vendor management, and cybersecurity protocols tailored for AI systems are also expected. The core principle is to ensure that for any decision that materially affects a customer or financial stability, there is meaningful human intervention and the ability to override the machine.
