The Securities and Exchange Board of India (Sebi) is set to undertake a comprehensive overhaul of the small and medium enterprise (SME) listing framework, while also tightening accountability around the use of artificial intelligence (AI) and machine learning (ML) in financial markets.

Speaking on the sidelines of the Ficci Annual Capital Markets Conference in Mumbai on Wednesday, Sebi Chairman Tuhin Kanta Pandey said the regulator was examining several shortcomings in the SME framework, including difficulties in trading, market making, underwriting and migration to the main board.

The regulator has seen odd lots being created on the SME platform, making it difficult for investors to trade. Measures aimed at controlling retail participation, such as increasing trading lots and application sizes, have also not delivered the intended outcome, Pandey said.

The market-making framework was another area of concern, with the chairman saying it was not working properly and was adding to costs for SMEs. The underwriting system also needed a rethink, while requirements linked to migration and paid-up capital needed to be reconsidered.

“We will bring a comprehensive reform proposal, a consultation paper will be issued for that,” Pandey said.

On AI and ML, Sebi is preparing guidelines for responsible use of the technology in the securities market. Pandey said an earlier consultation on AI had not yet been concluded, but developments since then had prompted the regulator to incorporate newer issues into the proposed framework.

The focus, he said, would be on facilitating the use of AI while ensuring that responsibility and accountability are clearly defined.

“Technologies are there, and we will certainly help in AI technology to be facilitated, but then we must have responsibility and accountability more clearly outlined,” Pandey said.

In his speech, Pandey said every Sebi-regulated entity would remain fully responsible for any AI or ML tool it uses, whether developed internally or procured from a third party. This responsibility would extend to the privacy, security and integrity of investor data, as well as the outputs generated by such systems.

Sebi’s proposed guidelines will follow a tiered approach, with clear accountability and governance controls. They will require safeguards such as “kill switches” and “humans in the loop”, alongside data controls, to balance technological innovation with investor protection.

The regulator has already deployed AI through Project Sudarsan and R(AI)DAR to identify suspicious financial promotions and potentially misleading advertisements. It has also created the Cyber Suraksha Portal to strengthen information sharing and cyber resilience across the market ecosystem.

Sebi is also reviewing the Listing Obligations and Disclosure Requirements (LODR) and delisting frameworks.

Further, for real estate investment trusts (Reits) and infrastructure investment trusts (InvITs), Sebi is examining whether to offer greater flexibility for investments in under-construction projects within certain limits.