SEBI Develops Framework for AI-Based Trading Oversight

NewKerala and ANI report that the Securities and Exchange Board of India (SEBI) is preparing a framework of guidelines for AI-driven trading to address efficiency and rising cyber risks. NewKerala's account says Chairman Tuhin Kanta Pandey warned at a press event about cyber threats that could threaten market integrity, and the coverage names an initiative called Project Jagrook, described as an AI-enabled investor awareness programme. BusinessLine snippets and ANI reporting additionally note Pandey framed foreign portfolio investor (FPI) outflows as part of normal global investment cycles and urged asset management firms to channel investors toward regulated financial products.

Regulatory interest in AI-driven trading typically centres on controls that reduce operational and security risk without unduly constraining latency-sensitive strategies. For practitioners, that pattern implies a likely emphasis on demonstrable patch management, access controls, secure model deployment, and incident-detection telemetry-areas that exchanges and regulators commonly require when algorithmic strategies interact with market infrastructure.

Industry reporting places SEBI's move alongside global conversations on algorithmic supervision and third-party model risk. Observers in other jurisdictions have been testing requirements for model explainability on high-frequency strategies, resilience testing under adversarial conditions, and vendor governance for third-party ML components. Those precedents frame the kinds of compliance expectations Indian market participants may need to prepare for.

Follow SEBI's formal consultation documents and any timelines published by the regulator, as those will define scope (retail versus institutional, onshore versus offshore firms), technical requirements, and phased compliance windows. Also monitor whether Project Jagrook issues public materials or APIs that change retail exposure to automated advice, and whether exchanges publish technical standards for secure model integration.