SEBI: AI for oversight, not for decisions

Tuhin Kanta Pandey, Chairman of India's Securities and Exchange Board (SEBI), has clarified that artificial intelligence (AI) can be leveraged for financial market supervision, but its role in making regulatory decisions should be constrained.

Speaking at the Global Fintech Fest 2026, Pandey emphasized that an AI-generated alert is not a finding in itself, and that enforcement and adjudication cannot be delegated to "black boxes." SEBI is already using AI for functions like cybersecurity compliance and social media monitoring, but stressed that ultimate regulatory decisions must retain human oversight.

SEBI aims to use AI to make supervision more predictive. According to Pandey, AI can process large datasets, detect patterns, and direct supervisory attention to emerging risks. This allows for a shift from periodic checks to continuous, near real-time monitoring, often conducted off-site.

However, Pandey cautioned that mere human presence in the decision-making loop is insufficient, as humans can overtrust machines. He highlighted the need for competent, empowered, and accountable oversight. Safeguards must cover AI models and their deployment, including explainability, validation, continuous testing, data quality, and bias controls.

Pandey also warned that AI could create concentration risks in financial markets if multiple entities rely on common models or cloud infrastructure, which could become shared points of failure. Regulatory responsibility cannot be outsourced to technology providers, and critical technological dependencies must not become regulatory blind spots.