The Securities and Exchange Board of India (SEBI) plans guidelines on the use of AI in capital markets, chairman Tuhin Kanta Pandey said at the opening of the regulator’s pavilion at the 45th India International Trade Fair on 15 November. The proposed guardrails would cover disclosures, the accountability of regulated entities and human oversight of critical decisions, BFSI Network reported. SEBI regulates India’s securities markets, and the fair was organised by the India Trade Promotion Organisation.
Key points
- Pandey said SEBI would issue detailed guidelines for responsible AI use in capital markets.
- The proposed guardrails cover disclosures, regulated entities’ accountability and human oversight of critical decisions.
- SEBI consulted in June 2025 on principles for AI and machine-learning applications in securities markets.
- Separately, SEBI introduced a cyber-incident response mechanism in May 2026.
Pandey puts people in critical decisions
Pandey said the guidelines would set out how AI could be used responsibly in capital markets while protecting investors, The Hindu Businessline reported. His proposed human-in-the-loop approach would keep critical decisions under human oversight rather than leave them entirely to automated systems. That places responsibility with the regulated organisation even where an AI application contributes to a decision.
The three elements Pandey identified address different points in an AI-assisted process. Disclosure concerns what a regulated entity makes visible about its use of AI. Accountability concerns which entity remains answerable for that use. Human oversight concerns the decisions that cannot be handed wholly to a system. Pandey presented them as subjects for forthcoming guidelines, rather than as obligations already in force under the announcement.
For market participants, the distinction matters. A proposal for human involvement in critical decisions identifies a limit on automation, but the eventual guidelines will determine how that limit operates. Pandey’s account of the planned guardrails put investor protection alongside the use of new technology, rather than treating the introduction of AI tools as a sufficient reason to remove people from consequential decisions.
SEBI’s June 2025 consultation set the groundwork
SEBI put forward a consultation paper in June 2025 proposing principles for the responsible use of AI and machine-learning applications and models in securities markets. Its stated aims were to retain the benefits of those applications while limiting risks to investor protection, market integrity and financial stability. Pandey’s announcement moves the discussion from proposed principles towards detailed guidance for the entities SEBI regulates.
That progression gives the planned guidelines a narrower institutional setting than a general debate over AI safety. SEBI’s concern is with applications in securities markets and with organisations already subject to its supervision. Disclosures and accountability would therefore bear on firms using the tools, while the proposed human-in-the-loop approach would bear on how they make critical decisions. The consultation’s three stated concerns — investors, market integrity and financial stability — explain why those uses fall within the regulator’s remit.
India is also considering AI rules beyond securities markets. AI Affairs reported plans for a separate consultation on AI safety and deepfakes. SEBI’s planned guidelines address a different group: participants in capital markets that answer to the securities regulator. The same technology can consequently be examined through a market supervisor’s duties as well as through a wider government discussion of safety.
SEBI’s May mechanism addresses cyber incidents
Pandey also described work elsewhere in SEBI’s technology supervision. In May 2026, the regulator introduced a structured mechanism for responding to cyber incidents, aligned with the Financial Stability Board’s FIRE reporting template. The mechanism provides a standardised way to report and manage such incidents across the securities sector. It addresses cybersecurity rather than serving as the AI guidance Pandey said SEBI would issue.
Among the threats Pandey raised were attacks on financial institutions and cryptocurrency platforms, and digital fraud involving social media influencers. He said SEBI was strengthening its cybersecurity framework to address them. In the same account of the regulator’s preparations, he said SEBI was taking an inventory of cryptographic systems to identify vulnerabilities associated with quantum computing. He expected quantum-related risks to become more pronounced globally around 2032–35.
SEBI is developing technology for its own supervisory work, too. Pandey said the regulator had developed AI tools including Sudarshan, which is being demonstrated at international regulatory forums. He also described work through international initiatives and supervisory toolkits to strengthen SEBI’s AI capabilities. Those uses put the regulator on both sides of the proposed guidelines’ subject: overseeing AI in the markets it supervises while building AI into its own supervision.
