Go to Contents Go to Navigation

(Yonhap Interview) AI yet to conquer stock analysts, won't 'take our jobs'

All News 15:36 March 14, 2016

By Kang Yoon-seung

SEOUL, March 14 (Yonhap) -- With Google Inc.'s self-learning program posting a series of victories over the world's top Go player, industry watchers from home and abroad cast a brighter outlook over the segment, with some predicting that the new era for artificial intelligence (AI) has just begun.

The groundbreaking win of AlphaGo against South Korean Go player Lee Se-dol drew attention from all segments, from the tech world to the medical sector. Some expect AI technology to eventually drive many people out of their jobs. The financial industry is no exception.

South Korea-based financial firm Summit International Capital, which has been focusing on adopting AI technology in securities trading, however, told Yonhap News Agency this will not be the case for the industry, at least for the time being.

"Unlike the game of Go, the securities market focuses on predicting the future," said Kay Lee, one of the key developers of the robo-adviser program dubbed "Q-BO." "Go is played on a limited board, where black and white stones are placed equally. In Go, it's comparatively easy to predict what's going to happen."

Go, known as "baduk" in Korea, originated in China more than 2,500 years ago. It involves two players alternately putting black and white stones on a checkerboard-like grid of 19 lines by 19 lines. The object is to claim larger territories than one's opponent by surrounding empty areas of the board with one's own stones.

"As AlphaGo can analyze the gameplay from countless games, it can predict what will happen. But this won't happen in securities investment," Lee added.

Lee has been making efforts to develop the so-called robo-adviser program that takes investors' personal tastes and strategies into consideration. Summit International Capital is one of the official partners of KDB Daewoo Securities Co. and is currently in talks with other brokerage firms.

"The AI technology refers to making a program think and learn by itself to achieve a certain result. What we are aiming for is to apply this to the financial market," Lee said. "Although nobody, including us, has reached that level for now, we are getting closer."

Robo-adviser programs, such as Q-BO, distinguish themselves from traditional program-based trading ones as they can apply far more complicated algorithms. They can implement more sophisticated investments based on foreign exchanges, oil prices and key rates, while current programs merely repeat buying and selling at predetermined prices.

Q-BO boasts different algorithms to suit the different needs of various investors' appetites, each focusing on a different basket of shares. Amid the rising attention of the AI industry, brokerage houses from home and abroad have been making efforts to tap deeper into the segment.

Market watchers not only expect robo advisers to bring changes to institutional investing but to individuals as well.

Citing industry data, Summit said the global market for robo advisers stood at US$14 billion in 2014 but will eventually grow to $255 billion in 2019.

Although the current status of robo advisers cannot be considered to be "self-learning," nor enough to catch up to the minds of analysts, Lee said the new system can make trading much easier.

"If the previous method of finding a profitable investment portfolio was like finding a needle in a haystack, robo-adviser programs are like magnets," Lee said. "They can handle problems in a simplified manner that up until now have required numerous people."

But when it comes to the AI industry, it is inevitable to ask the question, "will it take our job?"

Lee said robo advisers will continue to serve as assistants for analysts and their development does not necessarily mean that they will replace humans, citing a story where fund managers and monkeys battled to pick more profitable portfolios, and the humans lost.

"The story indicates the complexity of securities investment. But it is clear that robo-advisers can reduce human error," Lee said, adding the new systems will assist humans rather than replace them. "This is to make finding needles easier. It is about efficiency."

Market watchers also cast questions over the reliability of
robo advisers' algorithms as they do not have face-to-face meetings with investors to assess their investment appetites, while the systems also have not been verified.

Thus, the first batch of robo advisers in the market will continue to play a limited role for the time being.

"As for AlphaGo, the program's actions are based on games previously played by humans. Robo-adviser programs will also have to study human data. But unlike Go, this does not mean that they will be able to predict the future," Lee said. "There are numbers that can only be made by humans through offline research and onsite observations."


Send Feedback
How can we improve?
Thanks for your feedback!