Will AI change banking as we know it?
The banking sector is poised to follow the artificial intelligence (AI) revolution. But at what cost?
The Reserve Bank of India’s (RBI) recent committee report and framework titled “Framework for Responsible and Ethical Enablement of Artificial Intelligence (FREE-AI)” is a testament to imminent AI integration in banking. But organisational adoption of AI has gained infamy for layoffs and mixed opinions on user experience.
According to the World Bank, 89% of adult Indians had a financial account in 2024. This was a huge leap from 35% in 2011. Digital payment adoption and the Pradhan Mantri Jan Dhan Yojana (PMJDY) – a financial inclusion program launched in 2014 by Prime Minister Narendra Modi – are credited for motivating such a shift. With that many citizens connected to the banking sector, even a minor change will have a huge impact. But experts suggest that the banking sector will adopt AI slowly, cautiously, and in a way that benefits all.
AI Integration in Banks: Inevitable, but Time Consuming
“There are many risk factors in AI right now, which concern three primary factors: security, trust, and accountability. These risks must be dealt with before putting AI in any decision-making position. But it may very well be used by human employees as a tool in the meantime. RBI’s FREE-AI Framework too considers, at large, those three factors, among various others,” Deepak Kumar, (retd.) executive director of RBI, told HT. At present, Kumar is the director of RBI’s Institute for Development and Research in Banking Technology.
“About security, what the banking sector could do with AI, the criminals could do as well. So, new AI-powered cybersecurity abilities must be considered. About trust, if an AI ever proposes a decision, it’ll be probabilistic with inherent randomness. But the data, model, and ‘runtime control’ (control over application /function executions) provided to the AI determine how that probability was calculated, i.e., a justification; That justification must win the trust of its users before a go-ahead is given. About accountability, it is not only a legal concern, but it also plays a huge role within the hierarchy of the bank, especially when Agentic AI may make decisions. Therefore, accountability must be drawn out before adopting AI,” explained Kumar.
When asked about the probability of AI taking the jobs of bank staffers, he said, “A human presence is very important. Still, how many banking jobs may be lost to AI in the future cannot be predicted right now; nor can we predict when the jobs will be lost.”
But Kumar insisted that AI research must continue, even within the banks. “Unlike other sectors, we in the banking sector must continue rigorous research on AI security before adoption in the business functions. Research, development, and innovation must be calibrated for an effective [AI] adoption. This requires reskilling and upskilling to keep pace. And like every revolution, there will of course be initial hysteria and fear about employment and user-experience.”
AI Will Help Both Banks and Customers
According to a 2021 report by the Pew Research Center, India has roughly 1.2 billion lower-income individuals. Therefore, banks in India cannot and do not limit themselves to a well-off clientele. Catering effectively to low-income customers is crucial.
“Especially in banks that cater to customers with low incomes – such as farmers or small-scale business owners – AI models can be of substantial help,” said Ramakrishna Regulugedda, former chief general manager of the National Bank for Agriculture and Rural Development (NaBARD).
“There are AI models that are helping banks study, with the customer’s consent, non-traditional data such as utility bill payments, e-commerce behavior, or even mobile usage patterns of new-to-banking customers who lack a credit history. This secures the bank by helping it make better decisions on lending money. And it also helps the customer in interacting with multiple AI-enabled lenders/banks by recommending to the bank the optimum product/service for the customer’s query. In other words, getting the customer a better deal,” justified Regulugedda.
When asked about possible layoffs and AI employees, he said, “I don’t see AI taking over the banking sector, nor can I envision any layoffs in the sector due to AI. This is primarily because trust plays a huge role in banking, and machines can’t provide trust.”
Why Does Trust Matter?
The idea of a “human-free” bank isn’t new, believes Dr. Vikas Singh, adjunct professor at the Indian Institute of Public Administration, and co-founder of Crux Management Services Pvt Ltd.
“For years, technologists have imagined a “human-free” bank, with algorithms approving loans in seconds, chatbots handling customer angst, and automation taking care of the rest. Machines have indeed conquered the routine: KYC checks, transaction monitoring, fraud alerts. Yet the notion of an entirely automated bank remains a convenient fiction,” expressed Dr. Singh. “Banking has never been only about processing money; it has been about earning trust.”
He believes that AI can’t bond with humans nor understand human emotions, which undermines its potential as a banker. “When uncertainty strikes, whether through job loss, debt, or bereavement, clients seek understanding, not efficiency.”
“In India, the shift to digital banking has been swift, yet seven in ten customers still prefer visiting a branch when problems arise. Trust, after all, is easier built across a desk than across a screen. Centuries of relationship-based banking and a cultural preference for personal contact keep the human banker indispensable. Even in a high-tech age, the most valuable connections remain human,” he said.
“Corporates are beginning to realise that the novelty of a chatbot or a non-human interaction is subtracting value. The customers want to talk to real human beings. That’s a push the corporates need to handle more delicately,” he concluded.
