RBI Deputy Governor Calls for Responsible AI Integration in Finance

AI promises efficiency and inclusion but needs strong oversight: RBI Dy Governor Sankar - The Economic Times

Artificial intelligence (AI) is transforming industries with unprecedented speed, but it also poses serious risks if left unchecked, said Reserve Bank of India (RBI) Deputy Governor T Rabi Sankar on Tuesday. Addressing the Global Fintech Fest (GFF) 2025 in Mumbai, he cautioned that while AI promises extraordinary efficiency, inclusion, and innovation, it must be accompanied by rigorous oversight to safeguard financial stability. "AI is demonstrating its ability in ways unimaginable a few years back. India, too, has been a notable participant in this journey. But as with all powerful innovations, AI carries a dual narrative of promise and peril," Sankar said, speaking at a session on 'Responsible AI for Finance: Balancing Innovation with Financial Stability' at Global Fintech Summit 2025 at Mumbai. He stressed that the integration of AI into financial systems must be treated with profound responsibility. "In finance, the margin for error is narrower, as financial institutions are built on trust and economies prosper on stability," he said. RBI Deputy Governor added, AI in financial services should be designed with safety at its core, not as an afterthought. "AI systems must be subjected to rigorous oversight and layered with inherent checks. Financial AI applications, in particular, must not be allowed to destabilize markets, payment systems, or consumer confidence. Safety must be built into the design from conception to real-world deployment. Retrofitting safeguards later can be inadequate and even destabilizing," he noted. Sankar underlined Central Bank's focus on ensuring that technological advancement aligns with financial prudence. Citing the example of RBI's own innovations, he spoke about 'MuleHunter.ai', a system developed by the RBI Innovation Hub to combat the growing menace of mule accounts and fraudulent accounts used for money laundering and cybercrime. "Unlike the traditional rule-based systems currently used, MuleHunter offers greater accuracy and precision with significantly lower false positive rates. The model has already been deployed in about 20 commercial banks," he said. Responding to concerns about job displacement due to AI, Sankar said that its long-term impact would depend on whether it follows the trajectory of past transformative technologies like the industrial revolution or the invention of electricity. "Whether AI displaces jobs or creates new opportunities will hinge on how we adapt and govern its use," he added. (ANI)

RBI Deputy Governor Warns Of AI's Threats, Talks Of Way Forward

Reserve Bank of India (RBI) Deputy Governor T. Rabi Sankar warned that Artificial Intelligence (AI) poses "unprecedented threats" to the financial system if left without effective control. Speaking at the Global Fintech Fest in Mumbai on October 7, 2025, Mr Sankar stressed that financial institutions must approach AI integration with "profound responsibility". Consequently, he called for mandatory "safety by design" checks throughout the AI life cycle to prevent market and payment system instability. Furthermore, he highlighted risks including weakened accountability, bias in training data, and the difficulty of explaining opaque models, which impedes regulatory scrutiny. The Deputy Governor also stated that financial AI applications must not destabilise consumer confidence or the economy. Consequently, he presented five guiding principles, named the 5Ts, to ensure that AI remains a force for good: Trust, Transparency, Training, Technology for Good, and Togetherness. Notably, this comes after another RBI Deputy Governor, M. Rajeshwar Rao, spoke on AI in finance, urging measured and responsible adoption in the banking sector in September. According to Rabi Sankar, the power of AI in finance can expand financial access, strengthen safeguards, and reimagine efficiency. At the core, AI-driven solutions facilitate better credit assessment by using alternative data, such as utility payments, for customers without bank accounts. Similarly, the ability to process massive data sources can help institutions make real-time detection of fraud through identifying unusual transaction patterns. Furthermore, AI provides a paradigm shift in operational efficiency, achieving cost reduction in back-office processes, Know Your Customer (KYC) compliance, and loan processing. It also delivers 24/7 customer support via virtual assistants, and makes investment advice affordable to small investors through 'robo-advisors'. Elsewhere, AI-driven regulatory technology applications help regulated entities with better compliance outcomes. However, despite these benefits, this technology presents several risks. AI systems trained on historical data are likely to perpetuate or amplify existing discrimination in areas like credit profiling. The Deputy Governor said that: "Even small biases in training data can lead to systematic exclusion of population groups from accessing financial services." Also, algorithmic opacity makes identifying these biases difficult. This "Black Box" problem, or lack of explainability, makes models non-transparent, undermining accountability expected from both regulators and auditors. Additionally, the widespread use of AI-driven trading models may introduce systemic risks, such as herding behaviour, where multiple models could align their output with a wider consensus and amplify volatility. Likewise, AI misjudgments possess the capacity to trigger market dislocations. Ethical concerns also arise from using "behavioural data for manipulative cross-selling or risk profiling". Crucially, over-reliance on automation risks losing oversight, resulting in delayed intervention when problems occur. Elsewhere, Rabi Sankar noted that legal frameworks often struggle to keep pace with rapidly evolving technology. Rabi Sankar directly addressed how authorities must enable innovation while simultaneously safeguarding systemic stability. This balance acts as a necessity: it ensures AI reinforces the financial system rather than undermining it. Indeed, rigid regulatory frameworks actively dissuade experimentation, potentially reducing AI to a tool only deployed by the largest market players. Conversely, unbridled adoption, especially in high-impact areas, creates vulnerabilities that remain invisible until they "snowball into crises". Authorities must therefore actively encourage innovation, promoting policies that create safe spaces for experimentation, such as sandboxes. Furthermore, they must facilitate open digital infrastructures and provide access to quality data, enabling firms to innovate with confidence. The system also requires incentives so firms view governance "not as a burden but as a competitive advantage". Additionally, the financial system demands the highest degree of prudence, compelling institutions to ringfence critical infrastructures from unchecked risks arising from untested deployments. Thus, the industry must institute practices like stress-testing models in diverse scenarios, red-teaming to identify vulnerabilities, and adopting explainability tools. Furthermore, Rabi Sankar stressed that the industry must also develop its own ethical guidelines and self-regulatory codes to complement regulatory oversight. Ultimately, AI applications must be designed so they cannot inadvertently destabilise markets or consumer confidence. Rabi Sankar pointed out that embedding AI in finance demands continuous research, experimentation, and learning: as models, techniques, and risks rapidly evolve. Therefore, the financial system requires extensive partnerships, compelling "industry, academia, regulators, and start-ups" to collaborate and co-develop solutions. Furthermore, institutions must implement systems for responsible data governance, ethical data sourcing, and privacy-by-design in every model. They must also develop common standards, toolkits, and disclosure mechanisms so that model design, training data, and decision logic "can be explained to regulators and customers". Consequently, internal policies require revision to embed AI risk assessment directly into the product lifecycle. Looking ahead, the Deputy Governor stated the road for responsible AI demands a phased approach, deliberately balancing stability, innovation, and inclusion. Crucially, the human element remains central, and thus, the coming decade requires a parallel focus on AI literacy for individuals "to engage confidently and safely with these new tools". Meanwhile in the short term, the focus involves awareness and capacity building: financial institutions must train personnel and strengthen internal governance structures. In the medium term, AI will begin to play a substantial role in credit decisioning, and financial inclusion, and others. Ultimately, in the long-term India "can aspire to become a trusted global hub for responsible AI in finance". Moreover, by showcasing how innovation can coexist with strong safeguards, India "can set an example for emerging economies and the Global South". The RBI's focus on AI holds significant importance because financial institutions actively deploy this technology, making it a current operational reality, rather than a future problem. And two RBI Deputy Governors speaking on the subject in quick succession highlights the central bank's focus on this technology's possibilities as well as challenges. Importantly, the RBI warns that if the technology remains unattended, it can pose unprecedented threats to a financial system built entirely on trust. Therefore, the central bank set up the FREE-AI Committee, which articulated a set of guiding sutras (foundational principles) for ethical AI adoption. Specifically, the RBI is stressing that institutions adopt safety-by-design checks and train personnel to strengthen governance and trust. Ultimately, this preparedness proves necessary to ensure AI reinforces systemic stability in the Indian financial sector rather than undermining it.