Stern students, watch out: if NYU doctoral candidate Spencer Greenberg has his way, days of human investing may be numbered.
Greenberg’s project, which he calls Artificial Intelligence Learning Technology Investment Software, is able to comb through colossal amounts of financial data faster than any human being could. As it analyzes the different styles of investing that have worked best in decades past, it modifies itself accordingly for future use.
“Humans will always be better at certain tasks, but there are other tasks that the computer is going to get better at,” Greenberg said.
As more data is fed into the software over time, it will generate more successful trading strategies by itself. Think of the way Netflix works: The more movies you rate online, the better the recommendations Netflix is able to provide. That’s because Netflix will “learn” your movie taste based on the previous data (i.e. the movies you rated).
“Computer[s] can actually get smarter and better at what they do,” Greenberg said.
But some people don’t buy the idea that computers can trade stocks in place of functional humans.
Stern graduate Cathy Gao, who had studied Greenberg’s ideas in class, is skeptical about the software’s reliability.
“I think that you can’t anticipate the market fluctuation even with technology like machine learning,” she said. “Especially in any economy like this.”
And while Stern professor Dennis Shasha is not completely opposed to the idea of Greenberg’s A.I. software, he is wary of the idea of a human-less financial future.
“I agree in principle that a human/machine learning strategy is the way to go,” Shasha wrote in an e-mail. “But the human is important to choos[ing] the strategy and to manag[ing] risk.”
See the original version, published on September 16, 2010, at the WASHINGTON SQUARE NEWS.