To Better Detect Market Manipulation, NASDAQ Starts Using AI-Powered System

10/11/2019

NASDAQ is the world's largest stock market by volume. To eliminate fraudsters, it needs to constantly improve its systems, as well as to always monitor for any possible illicit attempt in tricking its analytics capabilities.

These include, and not limited to: manipulations to inflate a stock’s closing price; people in rapidly buying and selling stocks to create false activity impressions; and instances where people place large buy or sell order with no intention of actually executing.

Working alongside its Machine Intelligence Lab and the U.S. market surveillance division, NASDAQ's monitoring capabilities have been improved with AI.

According to NASDAQ, the system has been improved with deep-learning algorithms which work in tandem with human analysts to keep watch over its roughly 17.5 million daily trades.

NASDAQ
NASDAQ's stock market display at Times Square, New York

Tony Sio, head of marketplace regulatory technology at NASDAQ said that:

“Working closely with our U.S. surveillance team, we are able to train our models based on their experience in monitoring data directly from the trading engine of the world’s most liquid and dynamic market."

"Through transfer learning, we have now built a framework to provide those learnings to other marketplaces around the world. We believe this is a major step in the evolution of how we use artificial intelligence technology to maintain the integrity of capital markets."

The AI works by augmenting existing software analysis systems that use statistics and rules to flag signs of market abuse.

With the ability to accurately identify patterns, the AI system should help the overall system in detecting complex methods of frauds, which in turn can reduce the burden of its human analysts.

To make this possible, the AI was trained to detect particular subsets of abuse by learning from historical examples.

By understanding the patterns of abuse, the AI should be able to detect similar suspicious activity whenever it sees one. And when it does, the system will alert a human analyst with the appropriate expertise.

According to Michael O’Rourke, head of machine intelligence at NASDAQ:

"AI and machine learning have broad application across our company – from predicting market trends with Nasdaq’s proprietary data or creating more sophisticated market surveillance capabilities."

"We want to make every one of Nasdaq’s businesses better and smarter by providing them with richer data and more context-specific information. AI is going to play a crucial role in building the next generation of technologies for not only the capital markets, but every industry."

With the human analyst investigating and returning the outcome back to the system, the deep learning algorithms should improve continuously by refining its understanding. And also because it uses AI, the system can also be trained with different types of abuse, to learn about newer ones.

In other words, the system should be able to adapt to new patterns as fraudsters’ tactics evolve and become more sophisticated, as explained by Tony Sio, "the patterns and types of abuse that are happening are constantly evolving as well.”

But since AI can be fooled, Doug Hamilton, NASDAQ's managing director of AI, said that NASDAQ is initially rolling out the AI-powered system on top of the old one, rather than replacing it immediately and entirely.

Having human analysts working in tandem with the AI should add an additional layer of insurance, he explained.

If the system is a success, NASDAQ plans to roll it out globally.