The Securities and Exchange Board of India (SEBI) plans to issue detailed guidelines governing the responsible use of artificial intelligence within capital markets, according to a report by PTI published by MediaNama. SEBI Chairman Tuhin Kanta Pandey made the announcement at the India International Trade Fair in New Delhi on Saturday. The regulator aims to establish specific guardrails for artificial intelligence deployment, focusing on mandatory disclosures, accountability for regulated entities, and a requirement for human oversight in critical decision-making processes. PTI’s report does not specify a timeline for the release of these guidelines.
These upcoming rules build upon a consultation paper released by SEBI on June 20, 2025, which sought public comments until July 11, 2025. That earlier document proposed that firms using artificial intelligence in operations affecting clients, such as algorithmic trading or portfolio management, must inform investors about the model’s features, purpose, risks, and limitations. The consultation paper also suggested that third-party artificial intelligence services should be deemed as provided by the market participants themselves, ensuring that regulated entities remain legally responsible for compliance. SEBI already holds firms accountable for the AI tools they use, a stance reinforced by amendments made in December 2024 regarding data privacy and output responsibility.
Pandey emphasized that human oversight must be competent, empowered, and accountable, rather than merely symbolic. He previously noted at the Global Fintech Fest that humans can overtrust machines and that an artificial intelligence-generated alert is not a finding. The proposed framework requires firms to designate senior management to oversee artificial intelligence models, establish fallback plans for failures, and appoint independent teams to audit model accuracy. These audit findings and accuracy results must be shared with SEBI. Additionally, firms are required to test models in environments separate from live systems and retain input and output data for at least five years.
The regulator is also addressing broader systemic risks, including the potential for deepfakes to manipulate market prices and the dangers of firms relying on a small number of generative artificial intelligence providers. SEBI has developed supervisory tools such as Sudarshan, which scans public social media content for investment fraud and has flagged more than 20,000 instances of fraudulent content since November 2025. Pandey also highlighted growing concerns over cyberattacks and directed regulated entities to overhaul cybersecurity measures in May, aligning with a new incident response mechanism.