Addressing the valedictory session of the Global Fintech Fest (GFF) 2026 in Mumbai, Ms. Sitharaman called on the fintech industry to build systems that are “fast yet accountable, autonomous yet reversible, and innovative without diluting consumer trust”.
“AI may assist judgement, but responsibility must remain human and institutional,” she said, stressing that greater consequences from AI-driven decisions should be accompanied by stronger safeguards.
“The more serious the consequence of a decision, the stronger the requirement for review, explanation and appeal,” Sitharaman said. She urged regulators, boards and senior management to take direct responsibility for understanding how AI is deployed within their organisations.
“They must know where AI is being used, what decision it influences, what data it relies upon and what harm can arise if it fails,” she said.
Need greater scrutiny
The Finance Minister said AI systems involving higher-risk use cases should undergo greater scrutiny before deployment and throughout their operational lifecycle.At the same time, she cautioned against an approach that treats AI adoption and AI avoidance as the only alternatives. Institutions that remain dependent on legacy systems, she said, risk becoming less capable, losing competitiveness and eventually fading out.
“The relevant choice is therefore not between deploying AI and avoiding AI. The real challenge is between responsible adoption and well managed adoption,” she said.
New risks
Ms. Sitharaman also highlighted the dual nature of emerging technologies, particularly agentic AI and quantum computing, and questioned how financial institutions can balance breakthrough efficiency with growing systemic risks.She noted that tokenisation has reduced unnecessary intermediaries and enabled near-instantaneous asset transfers, while agentic AI has moved beyond providing recommendations to taking autonomous actions, compressing processes that previously took days into seconds.
However, she cautioned that the speed enabled by these technologies could amplify risks.
“Rapid transactions can cause software errors, fraud, and market shocks to spread just as quickly,” she said. While AI can accelerate fraud detection, it can also enable bad actors to conduct more sophisticated cyberattacks, she added.
“Quantum computing, meanwhile, could require a complete overhaul of traditional encryption standards” as its processing capabilities advance, Ms. Sitharaman said.
Sitharaman said the policy objective should be to avoid both excessive restrictions on innovation and the erosion of safety mechanisms in the pursuit of speed.
“Our goal should be neither to block innovation out of fear nor to let speed compromise safety,” she said.
Instead, the industry should distinguish between unnecessary operational friction and safeguards that are essential to financial stability, including human oversight and circuit breakers.
The Finance Minister also highlighted the possibility of creating a federated industry platform for India’s technology ecosystem.
Such a platform, she said, could provide the industry with a credible collective voice internationally and help Indian technology companies navigate global licensing and cybersecurity regimes through early engagement with foreign regulators.
