Technology developments have made market manipulation more complex and require the use of AI solutions for detecting and preventing market abuse.
Two Japanese brokerages – Rakuten Securities, a subsidiary of Rakuten Inc (TYO:4755), and SBI Securities, a subsidiary of SBI Holdings Inc (TYO:8473), are about to adopt artificial intelligence solutions, in order to adequately tackle evolving market manipulation schemes.
A report by Japan’s Nikkei says each of the brokerages will test AI software designed by NEC to detect market manipulation. The new systems are set to be ready for launch within the current fiscal year.
The AI software acts based on learning from past cases of illegal trading. It then flags suspicious activity. It searches for market manipulators who use techniques such as spoofing that involves placing large orders that are canceled before they are executed.
Brokerages have traditionally relied on human efforts to spot market manipulation. However, the volume and complexity of manipulation schemes have grown markedly. SBI Securities says it inspects 100,000 trading orders out of the 2 million placed daily.
An AI system, if implemented, would be responsible for both monitoring and inspection.
This is not the first time that Rakuten pushes into the world of AI. As reported by FinanceFeeds in April this year, Rakuten has teamed up with IBM Japan to create the “Rakuten AI Platform,” an internal Rakuten system which will launch chatbots with automatic response functions into customer support. The system was built utilizing APIs provided by IBM Watson.
Rakuten has been promoting AI research in areas like image recognition technology, as well as big data analysis.
Regarding the use of AI solutions to detect and prevent cases of market manipulation, we have to mention that early this year, Japan Exchange Regulation (JPX-R) and Tokyo Stock Exchange, Inc. (TSE) announced they are working on applying AI to market surveillance operations by using technologies developed by NEC Corp (TYO:6701) and Hitachi, Ltd. (TYO:6501).
NEC’s RAPID machine learning and Hitachi’s Hitachi AI Technology/H were used in preliminary investigations. The tests showed that AI solutions were highly accurate in identifying the possibility of unfair trading.