Sebi may turn to AI used in casinos to crack down on insider trading: Report

Sebi is considering the proposal as casino AI systems have shown high level of accuracy in raising alerts about abnormal trades.

CNBC-TV18 had earlier reported that the regulator has been utilising its database to identify front running based on coordinated price movements in equities purchased by funds, on the day of purchase.

The regulator had been alerted to such bad apples by market participants on various occasions. The big problem was to first spot the front running and then prove it.

Sources had told CNBC-TV18 that for the past few months, Sebi has been using artificial intelligence on its data trove to identify such synchronised stock moves on the days of purchase of stocks by funds.

The sources had also added that the regulator was likely to even broach the government to give it powers making it an enforcement agency under the IT Act once it is satisfied with its own algorithms and artificial intelligence software.

A number of casinos around the world as well as online casinos have already implemented artificial intelligence. AI is used to analyse the behaviour of players in the casino. They collect data and use algorithms to analyse why players prefer a particular game.

A report by New York Times said casino operators in Macau use hidden cameras and facial-recognition technology to track betting behaviour of gamblers. They also use poker chips with radio-frequency-identification technology and sensors on baccarat tables to collect data on a player’s performance and monitor inter-player collusion.

One of the functions of AI in casino is to help in risk management.  In this, AI is used to automatically analyse the bets being placed in the casino, check the modes of payments and build the player profile to raise an alert in case of suspicious trends. This can be used by Sebi to crack down on insider trading.

"The casino AIs come with superior risk management systems compared to the traditional data analytics that regulators use, because the margin of error for these AIs is very thin,” Economic Times quoted a source as saying. If casinos failed to detect even a single rogue gambler, they may have to pay the prize in seven figures, the source said.

For Sebi, it has become difficult to rely on human capabilities to monitor market participants who use advanced software. Till now, the capital market regulator depended on tips from exchanges and market intermediaries to probe white-collar crimes.