SEBI Chairman Tuhin Kanta Pandey has announced that the regulator is actively drafting a comprehensive framework governing the use of AI in capital markets. The framework explicitly supports AI deployment for fraud detection and investor servicing, but draws a firm line on autonomous algorithmic execution.
The centrepiece of SEBI's approach is a "Human-in-the-Loop" (HITL) mandate: no AI algorithm can hold final, independent control over trading execution or portfolio management decisions. A human supervisor must always remain in the control loop, with oversight authority over machine decisions.
SEBI's rationale rests on three pillars. First, the mandate prevents unchecked algorithms from triggering cascading market sell-offs—the flash crash risk inherent in machine-to-machine feedback loops. Second, it establishes human accountability: regulators can subpoena a compliance officer, not a piece of code. Third, it kills the 'black box' problem, ensuring AI models managing Indian wealth operate transparently without making biased decisions based on opaque parameters.
To anchor this framework internationally, SEBI is integrating standards from the International Organization of Securities Commissions (IOSCO), aligning Indian capital market rules with global best practice. This dual approach aims to encourage AI innovation while protecting systemic stability and investor protection.
For algo-trading firms and wealth-tech startups operating in India, the mandate significantly tightens compliance architecture. Firms must now design control systems, audit trails, and human approval workflows into every critical algorithmic decision.
