Artificial Intelligence In Stock Market: The Securities and Exchange Board of India (Sebi) has proposed a structured regulatory framework for the ethical and responsible use of artificial intelligence (AI) and machine learning (ML) by market participants. The use of AI and ML in financial markets has increased greatly in recent years, mainly due to better data availability, faster computing power, and improvements in tech infrastructure. Sebi said it noted that the rise of Generative AI and large language models (LLMs) has further opened new use cases for market participants.
Stock exchanges, brokers, mutual fund houses and other intermediaries are now using AI and ML tools, maily for advisory services, risk management, client verification, surveillance, pattern recognition, internal compliance, and cybersecurity, the market watchdog noted.
However, Sebi said that while these technologies can boost efficiency and outcomes, they can also bring certain risks if not managed properly.
“Their usage also creates or amplifies certain risks which could have an impact on the efficiency of financial markets and may result in adverse impact to investors,” Sebi said in the consultation paper.
To formulate the proposed guidelines for the use of AI and ML in the Indian securities market, Sebi had constituted a special working group. The working group studied the best practices of the use of these technologies in India and abroad and addressed concerns and issues related to it. The group also invited representatives from mutual funds, stock exchanges and brokers to understand how they are using AI and ML in their respective organisations.
Sebi has now invited stakeholders to submit their comments by July 11, 2025.
Sebi Guidelines For Using AI, ML In Securities Market
After multiple rounds of consultations, the working group finalised a set of guidelines for the same.
Here they are:
Set Up Internal Governance Teams
According to the proposed framework, market participants such as exchanges, brokers, mutual funds, and other intermediaries will be required to set up internal governance teams with “adequate skills, expertise and experience” to oversee the performance, controls, testing, efficacy, and security of the algorithms deployed.
A senior management, having appropriate technical knowledge and experience, is also required to be designated, who will be responsible for the oversight of the AI and ML based models.
“Since AI/ML applications can learn from live data and their model behaviour may hence change after deployment, market participants should conduct periodic reviews and on-going monitoring to ensure that the applications continue to perform as intended,” Sebi said.
Disclosure Norms For Using AI, ML
Market participants that use AI or ML for portfolio management, algorithmic trading, or investment advisory services will be required to disclose the same with their clients. Sebi also said that these participants will be required to clearly explain how these technologies are being used, how accurate the models are, the quality of the data behind them, the risks involved, and any fees that may apply.
Importantly, all of this information should be shared in simple, easy-to-understand language so that investors know what they are signing up for and can make informed decisions accordingly.
Guidelines For Testing AI, ML Models
Sebi has proposed market participants to thoroughly test all AI and ML models in the securities markets, not just before their deployment, but even after they are put to application.
The regulator has suggested a method called "shadow testing," where models are tested using real-time data without affecting actual operations.
Sebi has also proposed market participants to maintain proper documentation of all the models and record data for at least five years.
Tackling Bias Risk And Unfairness
Sebi also raised concerns about the risk of bias in AI systems. To tackle this, the regulator has urged market participants to actively check and correct for any unfairness in their models.
“Market participants should implement appropriate processes and controls to identify and remove biases from data sets,” Sebi said.
Data Privacy, Cybersecurity Guidelines
Since these systems rely heavily on collecting and processing large volumes of data, Sebi proposed market participants to devise clear and robust policies to protect that data.
Sebi has also made it clear that if there is a technical glitch or a data breach, market participants must report it without any delay.
