AI, a double-edged swordIf there was a financial market forecasting algorithm as smart and intuitive as AlphaGo that recently crushed the best living player in the game of Go, dubbed the most complex strategy board game, would everyone be rich? My quant, short for quantitative fund manager, did not hesitate to answer, “No.” Robo-advisers, or algorithmic advisers, are currently big on the local asset management market.
His explanation was simple. Such a smart machine would require huge investment, and few financial companies can afford it. Even if a brokerage house has such a sophisticated artificial intelligence program that can guarantee over 40 percent returns on investments, why would it spend time on customers for a fee of 5 percent? It would be busy making money for itself.
There are already many examples. In the United States, more than 70 percent of stock trading is already done through algorithmic or quantitative investing. Renaissance Technologies, founded in 1982 by James Simons, an award-winning mathematician at Harvard University, pioneered quantitative trading, now referred to as quant hedge funds that rely largely on computers, scientists, and mathematicians to guide investment strategies instead of gut-driven human managers.
Over the last three decades, the company was able to register an annual average return of 30 percent. It strictly serves member funds and is notoriously selective in accepting members. Most of the returns from the $27 billion hedge fund go to employees of the firm. Simons in 2014 was ranked 88th in Forbes’ list of the world’s billionaires, with an estimated net worth of $15.5 billion, far above Korea’s richest, Lee Kun-hee, chairman of Samsung Group, who was at 102nd.
Even an impeccable machine cannot promise over 40 percent returns. It would only return 40 percent over a single investment of 100 million won ($85,600). A client who assigns 1 billion won under its management cannot expect to make a similar profit from the remaining 900 million won.
The quant fund manager I talked to is one of the first in the local industry. He joined a financial engineering team launched by a securities company in 2005. The year 2007 was the industry’s heyday. Quant funds were in enormous demand due to the popularity of equity-linked securities. Experienced managers were paid multibillion-dollar annual compensations. The party ended with the Lehman Brothers bust in 2008. Quant funds began to lose money. Computer-based quant funds yielded negative 1.25 percent last year, hovering below stock funds run by human managers with negative returns of 0.45 percent.
The AlphaGo phenomenon revived the reputation of automated investing in the financial industry. Shinhan Investment and several other financial companies are putting together financial science teams. Quant managers are being scouted amid renewed demand for investment algorithms. Instead of stealing away or replacing their share, AI has added more human jobs to the industry.
The quant fund manager I spoke to predicted that a highly intelligent algorithm investing program would arrive in the financial scene within the next five years. The program would have strengthened intelligence to accurately place the timing and pricing to deploy and rebalance investments. But the machine would never replace the work of humans. No machine can invade the quintessential human realm of quick and cool judgment and gut drive required in moments of volatile stock trading. Even the deep neural networks powered by connected machines armed with self-learning graphics processing units cannot master unique human qualities.
Go guru Lee Se-dol managed to nab just a single win from the five matches with AlphaGo in the recent tournament. The last feat of the machine against human was the match between world chess champion Garry Kasparov and the IBM supercomputer Deep Blue. The 1997 match was the first defeat of a reigning world champion by a computer in a tournament game of chess.
But the trouble was that Deep Blue was useless outside the chess game. It scored zero on an intelligence quotient test.
AI scientists realized that computers can never entirely replace human intelligence. But science always advances beyond our imaginations. We cannot know how much further AI will go in emulating the human mind. Theoretical physicist Michio Kaku once said, “All science is a double-edged sword. One edge can cut against ignorance, poverty, and disease, the other side can cut against innocents.”
JoongAng Ilbo, March 24, Page 30
*The author is an editorial writer for the JoongAng Ilbo.