Summary of this article
AI can improve access, credit delivery and risk monitoring
Lack of safeguards may amplify bias, opacity and systemic risks
RBI outlines principles focusing on accountability, inclusion and trust
Artificial intelligence (AI) is set to transform the financial sector, but its adoption without adequate safeguards could amplify existing weaknesses and create new risks, stated RBI Deputy Governor Swaminathan J.
Speaking at a memorial lecture at SASTRA University in Thanjavur, he has stated that the focus should not only be on making finance more efficient but also on ensuring that it remains fair, accountable and inclusive.
AI Expands Access And Improves Efficiency
AI is already transforming the way financial institutions engage with customers, evaluate creditworthiness, risk management and compliance. Tools such as multilingual chatbots and voice-based systems can make financial services easier to access, especially for users who are not comfortable with formal documentation or English interfaces.
Regarding credit delivery, he explained that AI can complement the conventional approaches to examine larger trends like transaction behaviour and repayment flows. This helps in identifying borrowers with low formal credit histories, such as small businesses and first-time borrowers.
Fraud detection, risk management and security are the other areas where AI can be deployed to improve the effectiveness of financial transactions. He said large amounts of data produced during transactions and stored within banks, credit card networks and financial systems can help identify unusual transactions in real time.
Risks Of Bias And Data Control
However, he has warned that AI systems can reproduce and even amplify biases present in historical data. This can result in unjust results in other areas, like credit assessment, where decisions may seem objective, yet they may have underlying distortions. He claimed that this poses issues of consumer protection and inclusion.
Another concern is the lack of transparency in some AI models. He said financial decisions cannot remain unexplained, especially when they affect individuals’ economic lives. If a loan is denied or a transaction is flagged, institutions must be able to provide clear reasons.
Privacy of data is also a major concern. AI systems are based on data chunks, containing sensitive financial information. The Deputy Governor has also mentioned that institutions must have the right consent, storage and restricted access. The data governance, he said, should be considered a primary concern and not a secondary one.
Model risk was also highlighted, in which the imperfection of an AI system would have an impact on decisions made by a large group of users. When several institutions rely on the same models or suppliers, the risks could become interdependent, impacting the rest of the financial system.
Cybersecurity risks are increasing as well. While AI can strengthen defences, it can also be used by fraudsters to create more sophisticated attacks, including phishing and deepfake-based frauds.
Five Principles For Responsible AI Use
He presented five principles of responsible use of AI in finance. First, accountability should lie with the institutions, despite the possible support of AI systems in making the decisions. Second, systems should integrate fairness and explainability.
Third, institutions must adopt strong data governance practices covering collection, usage, storage and protection of data. Fourth, organisational capacity needs to be strengthened, including a better understanding of AI at the board and management levels.
For the fifth principle, he said inclusion should be a key objective. The impact of AI should be assessed based on whether it improves access to financial services for those who are currently underserved.
He said the effectiveness of AI in finance should be judged on three parameters: whether it advances inclusion, improves efficiency and strengthens trust.
Trust And Accountability
The RBI Deputy Governor said that technological progress in finance should not come at the cost of human judgment and relationships; banking continues to be built on trust, and innovation must support that foundation.
He added that while AI can improve efficiency and expand access, it must not reduce accountability or make financial systems less transparent.
