JP Morgan Develops AI that Executes Trades

Investment bank JP Morgan will soon launch its pioneer AI to execute trades in the bank’s European, Asian and US equities business.

JP Morgan told the Financial Times that its AI named “LOXM” has proven itself to be better compared to the traditional buying and selling method in a trial conducted in the bank’s European equities business in the first quarter of this year.

The job of JP Morgan’s LOXM is to execute client orders at high speed and at the best price. “Such customisation was previously implemented by humans, but now the AI machine is able to do it on a much larger and more efficient scale,” David Fellah, of JPMorgan’s European Equity Quant Research team, told the Financial Times.

Fellah stressed that the bank has had no risk management issues with LOXM. “The machine is restricted in its trading behaviour, as it learns under, and operates within, our general electronic trading risk framework, which is overseen by internal control groups and validated by regulators,” he said.

The AI developed by JP Morgan uses deep reinforcement learning (DRL) methods, which enable it to learn from billions of past trades, including real and simulated trades. In the investment bank setting, Fellah said that DRL’s potential uses include “automatic hedging and market making.”

Aside from learning from real and simulated past trades, the AI could one day be taught to understand individual clients so that the client’s behaviour and reaction could be considered when executing trades. Fellah, however, emphasized that any “customisation would only be if the client agrees to that.”