Illustration: Ashley Siebels
Of all the white-collar jobs in America, bankers, stockbrokers and lawyers have arguably had it the best: the most prestige, the most money, the most job security because of their extensive educations and personal knowledge.
Obviously, our future robot overlords would have to topple those privileged positions if they ever want to take over, right? And lo: despite the widespread conviction that automation is a risk only to blue-collar jobs like cashier and factory worker, the robots are coming for the finance workers and the lawyers just as fast.
Those jobs, featuring long days complicated thought and analysis, those long nights of moral judgments — well, the harsh truth is, a computer can do those better.
Exhibit 1 million in this case comes this week as BlackRock, a financial company that manages over $5 trillion in other people’s money, is trading in two dozen humans for much cheaper and more reliable algorithms. BlackRock is laying off about two dozen people, according to The Wall Street Journal. The reason: “The company has taken the view that it is difficult for human beings to beat the market with traditional bets on large U.S. stocks.”
It all comes down to profits. BlackRock manages about $278 billion in mutual funds that include the stocks of big American companies. Traditionally, humans have picked which stocks should be in those mutual funds, curating the portfolio so that it all rises in value regularly. The goal: beating the S&P 500 Index, the group of 500 companies that act as a proxy for the health of the U.S. economy and the stock market.
For decades, human stockpickers have had only one job: beat the S&P 500.
And in recent years, those humans have consistently failed.
Numerous studies have shown that “active” management by human stockpickers lags behind the “passive” investment of simply dumping money into an S&P 500 Index fund and leaving it there. In 2014, 80% of actively managed funds did not perform as well as the S&P 500. Investors have been pulling billions of dollars from funds managed by humans and pouring them into passive funds instead.
Wall Street’s love affair with algorithms
Enter the robots. Or rather, look around at the room, because robots have been making inroads on Wall Street jobs for nearly two decades.
It all started with decimalization in 2000, whereby stock prices were set by decimal places instead of the old method of using fractions. Decimalization made humans less relevant (and paid them less profit) so computers started to take over the New York Stock Exchange and fueled the rise of the all-automated Nasdaq. Hedge funds have long favored automated trading of stocks, rushing into high-frequency trading that executes trades in nanoseconds. The obsession for empowering computers in trading became so extreme that a popular strategy is “colocation,” in which hedge funds moved their actual servers closer to the New York Stock Exchange and other exchanges so the bits and bytes of trading wouldn’t even have to spend a few microseconds traveling down a longer length of cable. There are already nearly 1300 colocation centers in the U.S. — just giant gatherings of data servers, kept cool in dark, vast rooms. By one estimate, colocation itself may become a $50 billion market by 2020.
And as the computers do better and become smarter at trading, it has forced humans to become more thoughtful about what their jobs can be, besides just coming in to turn the switches on.
Increasingly, the pervasiveness of algorithms on Wall Street is even challenging the old advice-based professions, like mergers and acquisitions. Mergers and acquisitions bankers were once the leaders of Wall Street, paid millions of dollars to know which companies should combine, and to develop deep advisory relationships with CEOs to make it happen.
But here, too, human error played an enormous role: mergers more often fail than not, with estimates of that failure ranging from 50% or more, all the way up to 95% — a staggering proportion. Perhaps that’s why Goldman Sachs, one of the leading investment banks for mergers, decided to trade in a passel of 100 junior bankers for 75 coders who could match companies up by algorithm. (Goldman also replaced 600 human traders with 200 engineers as the bank’s chief financial officer crowed, “those 600 traders, there’s a lot of space where they used to sit.”)
Another financial function moving quickly under the oversight of artificial intelligence: financial advice. So-called “robo-advisors” from E*Trade, Fidelity and TD Ameritrade, Betterment and others help you structure your investment portfolio faster and cheaper than your financial advisor from Merrill Lynch or Edward Jones once would have. MyPrivateBanking estimates that robo-advice could grow to a $3.7 trillion market in the next three years.
The Lawyer Algorithm
Similarly, other white-collar professionals who give advice are facing competition from robots. In time for tax season, H&R Block struck a deal with IBM’s Watson to have a computer give tax-preparation advice on potential deductions.
IBM’s Watson, in fact, may also be behind the rise of what The Atlantic calls “Robolawyers,” computerized advisers who do for legal advice what TurboTax may do for taxes. The grunt work of junior lawyers, like diving into piles of legal documents, is ideal for computers. Automation of those unpleasant tasks may also force lawyers to spend more time on what matters: advice. When the research is easier to do for legal professionals, more time can be spent on making better arguments and coming to the right judgments. At least, that’s the theory.
No wonder some big-name technology brains are worried. Stephen Hawking, the physicist, is worried about artificial intelligence and what it could do to human society. Elon Musk, the brilliant CEO of Tesla, seems to fret regularly that computers will destroy us (even as he has ambitions to implant tiny computers into our brains.) Bill Gates, founder of Microsoft, thinks robots should have to pay taxes if they’re going to take over so many jobs. This kind of worry is why artificial intelligence is just now taking off. For 25 years starting in the late 1960s, years known as the “Dark Ages of artificial intelligence,” scientists held off on doing too much with the science for a variety of reasons: worries that it was all overblown, that AI would never scale up, and of course, pop culture feeding into fears that too much innovation too fast would lead to a robot apocalypse.
It can call sound pretty dark, but robots have their human defenders, too, including former Treasury Secretary Larry Summers, who is a staunch robot advocate (or robot apologist, depending on how you see it). The caveat from human allies like Summers is that artificial intelligence may not actually “take” human jobs, but merely help people by taking on grunt work and freeing us to do the really big thinking.
All we know is that, whether or not robots will ever truly be our overlords, they really are going after the best and biggest jobs first. Now that’s smar