Malhotra explained that AI is different from past technology shifts. While electricity enhanced energy and the internet improved connectivity, AI enhances intelligence, directly affecting how credit is priced and delivered. He noted that AI models using alternate data—such as cash flows and utility payments—could expand access to banking for many who lack traditional financial histories.
He highlighted India's unique advantage through its digital infrastructure, including Aadhaar, UPI, and other frameworks. AI, he said, could revolutionize financial judgment in a way similar to how UPI transformed financial transactions, making them more efficient and accessible. Immediate returns from AI could be seen in customer service, operational efficiency, and real-time fraud detection. Malhotra stressed the importance of AI enhancing human judgment and called upon banks to invest in technology, talent, governance, and partnerships.
He listed seven risks related to AI, including explainability, bias, concentration risk, data privacy issues, and accountability erosion. Malhotra emphasized that fairness in AI should be a design requirement, with banks ultimately responsible for their decisions.
He set immediate expectations based on the RBI’s committee recommendations, including creating a comprehensive inventory of AI models, formulating a board-approved AI governance policy, ensuring explainability in decisions impacting customers, and maintaining human oversight in potentially harmful situations. The RBI will continue to support banks through its regulatory sandbox and efforts to enhance digital tools, aiming to address gaps in underserved markets and prevent digital fraud through AI.