The Algorithmic Gold Rush
India's financial services sector is in the middle of a massive technological shift, with spending on AI expected to double in 2026. Banks, insurers, and fintech companies are embedding artificial intelligence into their most critical functions. AI models
now analyse creditworthiness for loans, power algorithmic trading on stock exchanges, detect fraudulent transactions in real-time, and manage customer interactions. This adoption is driven by the promise of unprecedented efficiency, the ability to process vast amounts of data, and the goal of expanding financial inclusion. With a robust digital public infrastructure including Aadhaar and UPI, India has a unique foundation for AI-led innovation in finance, allowing services to reach millions of people faster than ever.
When the Code Gets It Wrong
The rush to automate also introduces significant new vulnerabilities. One of the biggest fears is algorithmic bias, where AI models unintentionally perpetuate and even amplify existing prejudices, potentially leading to financial exclusion. If an AI model is trained on historically biased data, it might unfairly deny loans to deserving individuals or specific communities. Another major risk is the 'black box' problem, where the decision-making process of a complex AI is so opaque that even its creators cannot fully explain a specific outcome. This lack of transparency becomes a massive issue when an algorithm makes a mistake, creating an 'accountability gap' where it's unclear who is responsible for the error. Additionally, regulators and ministers have raised alarms about AI-powered fraud, such as deepfakes and hyper-personalised phishing attacks, which can operate at a scale and speed that traditional security measures struggle to handle.
The Regulator's Dilemma
Regulators like the Reserve Bank of India (RBI) and the Securities and Exchange Board of India (SEBI) face a delicate balancing act. They aim to foster innovation while safeguarding financial stability and consumer rights. In recent months, both have taken significant steps to create guardrails. The RBI has proposed a comprehensive framework requiring banks to manage risks associated with AI, demanding board-level oversight, periodic stress testing of models, and clear accountability for failures. For its part, SEBI has rolled out a mandatory framework for algorithmic trading, effective from April 2026, which requires every algorithm to be registered and assigned a unique ID to ensure traceability and accountability. These moves signal a clear shift from unchecked growth to governed innovation.
The Inescapable Human-in-the-Loop
The core of the current debate is the role of human oversight. As government officials and industry leaders have recently emphasised, AI should be a tool to assist, not replace, human judgment in high-stakes scenarios. Chief Economic Adviser V. Anantha Nageswaran has been vocal, warning that India cannot afford to wait for problems to emerge before acting. He and others argue that a 'human-in-the-loop' must be maintained, especially to prevent AI from becoming a tool for exclusion. This sentiment is echoed in the RBI's draft guidelines, which explicitly warn against 'automation bias'—the human tendency to over-trust automated outputs—and mandate meaningful human oversight for AI-driven decisions. A recent KPMG report found that a third of financial organisations are increasing human-in-the-loop oversight in direct response to AI-related concerns.
