SEBI to Issue AI Guidelines After Chief Flags Bias, Data Risks

India's market regulator will draft rules for responsible artificial intelligence use in capital markets, SEBI chief Tuhin Kanta Pandey said. AI tools can improve surveillance and fraud detection. They also carry risks around bias and data protection, Pandey added.

The guidelines will draw on international recommendations. Pandey did not specify a timeline. The announcement puts Indian brokerages, algo-trading firms, and exchanges on notice. Firms that use machine learning for trade execution, risk management, or client advisory will need to adjust compliance frameworks once the rules land.

SEBI already regulates algorithmic trading under its 2012 circular on co-location and algo order placement. The new AI guidelines would extend that framework to cover generative models, predictive analytics, and automated decision systems. The regulator has not said whether the rules will apply retroactively.

For now, the risk is uncertainty. Firms cannot model compliance costs or system changes without knowing the scope. A light-touch approach would leave current operations largely intact. Strict rules on model explainability or bias testing could force redesigns of proprietary trading systems.

Pandey's remarks came at a conference in Mumbai. He said SEBI will consult with market participants before finalizing the guidelines. The regulator also plans to coordinate with the Reserve Bank of India, which issued its own AI framework for banks last year.

No date has been set for the draft or the final circular.