Accountability for AI and ML outcomes in regulated entities

Securities and Exchange Board of India (SEBI), India's securities market regulator, recently amended its Intermediaries Regulations, provisions which came into force on 10 February 2025.

- ANURAG MIGLANI

The amendment defines "artificial intelligence and machine learning tools and techniques". The regulation essentially holds any entity (governed by SEBI) utilizing artificial intelligence (AI) and machine learning (ML) tools and techniques accountable for the outcomes generated by these technologies.

Since 2019, there have been reporting requirements regarding AI/ML systems. However, with this amendment, SEBI has effectively shifted the entire responsibility onto the users of such systems, Market Infrastructure Institutions and intermediaries like exchanges (National Stock Exchange (NSE), Bombay Stock Exchange (BSE), Multi Commodity Exchange (MCX) etc.), Stock brokers, Depository Participants (NSDL, CDSL etc.), Mutual Funds, Alternative Investment Funds etc.

But is this enough? The simple answer is no, it isn't.

Not to put it mildly, Indians like to gamble. The annual equity futures and options (F&O) turnover of NSE in India for the financial year 2024-25 (till 21st March 2025) is Rs 6,01,30,590.99 crores, i.e., Rs 601,305,909,900,000.00. And it's just equity contracts at NSE.

It does not include the F&O trading activity in other asset classes like commodity, currency, interest rates and activity on other Indian exchanges like BSE, MCX and others. Based on the volume of equity contracts traded, NSE is the world's largest derivatives exchange. Chicago Mercantile Exchange comes a distant second.