The guidelines, detailed in the consultation paper titled ‘Guidelines for Responsible Usage of AI/ML in the Indian Securities Market’, were shaped by a Sebi-appointed working group comprising experts who reviewed both Indian and international practices.
“AI and ML are becoming integral to how markets function, but their unchecked use could pose systemic risks,” the paper notes, adding that the proposed framework seeks to ensure that “innovation does not compromise investor trust or market integrity.”
What Sebi Is Proposing
The five-point regulatory framework focuses on:
- Model Governance
Entities using AI/ML must establish dedicated internal teams with technical knowledge to supervise and audit models. Senior management will be held directly responsible for the full AI lifecycle—from development to third-party oversight.
If AI/ML systems directly impact investors—such as in advisory services or automated trading—then clear disclosures must be made regarding model purpose, associated risks, accuracy rates, and limitations. Disclosures must use simple, investor-friendly language.
Sebi wants firms to conduct thorough model testing in simulated environments before real-world application. Additionally, all model documentation and logs must be preserved for at least five years to ensure traceability.
To prevent discrimination or skewed decision-making, firms are advised to use diverse, high-quality datasets. Staff must be trained to identify and correct biases in algorithmic outputs.
The proposed framework mandates strict data governance protocols, with a focus on robust cybersecurity and user data privacy. Firms must ensure compliance with prevailing data protection laws.
The regulator has clarified that AI/ML models used for internal surveillance or analytics will face lighter scrutiny, while those influencing investor decisions will be subject to tighter regulation.
“Our objective is not to stifle innovation but to ensure that any technology introduced in the market does not compromise investor safety or market stability,” said a Sebi official familiar with the development.
Why It Matters
AI and ML are already in use for functions ranging from real-time fraud detection to automated investment advisory. In 2024 alone, the number of brokerages using AI-powered customer services and predictive market tools grew by over 25 per cent, according to Sebi’s internal research. However, concerns have been mounting about the lack of accountability and transparency in model usage—especially when algorithms behave unpredictably.
What Happens Next
Public comments will be accepted until 11 July. After evaluating stakeholder feedback, Sebi is expected to finalise the regulatory framework later this year, making India one of the first major emerging markets to formally regulate AI/ML usage in its capital markets.
Market participants and industry bodies like the Association of Mutual Funds in India (AMFI) and BSE Brokers Forum are expected to submit their views in the coming weeks.