The future robo advisor: Smart and ethical?

Estimates suggest the number of robo-advice clients will climb to 17 million by 2021 from 1.8 million in 2016

Today’s robo advisors are expected to get a lot smarter. Tomorrow’s might even have a conscience.
 
So say those who study the use of artificial intelligence in finance—though a digital wealth advisor with morals is likely still a long way off.
 
The robot advisors of today use algorithms to provide low-cost, automated portfolio-management services to investors seeking discount advice. These easy-to-use tools have grown swiftly: Assets under the control of digital wealth managers are on track to surpass $1 trillion by 2020 after ballooning by triple digits annually since 2013, according to research firm Aite Group.
 
With their low fees and small minimum-balance requirements, robos have been a boon to middle-income investors, enabling them to access a service once reserved mainly for the affluent. But beyond recommending low-fee index funds based on factors such as an investor’s age, income and risk tolerance, and rebalancing portfolios as needed, the tools are limited in what they do.
 
Critics say not only are robots incapable of providing the kind of personalised, sophisticated financial-planning guidance that human wealth managers can deliver, they don’t understand right or wrong. So while robo advisors are required to put clients’ interests first when spitting out portfolio recommendations, they can’t truly act as fiduciaries, some observers say.
 
“A robot has no consciousness, no ethics,” says Vasant Dhar, a professor of information systems at New York University’s Stern School of Business who runs a robo advisor for institutional investors. For a robo to qualify as a fiduciary, “you’d have to have a machine that has a strong moral code and understands the implications” of violating it, such as a regulatory fine or lawsuit, he says.

 
Growing smarter
 
AI specialists say technology could help robo advisors develop some of the traits that differentiate human advisors from their robotic rivals. To get there, AI would have to accumulate vastly more data, while powering algorithms to simulate different scenarios and “keep introspecting,” Dhar says.
 
That will take time, if it happens at all, AI specialists say. But the idea is that eventually it will become much easier for an algorithm to provide fiduciary care, perhaps better than a human advisor, who might be tempted to recommend one product over another because of higher commissions or other incentives.

Meanwhile, robo advisors from pioneers such as Betterment LLC and Wealthfront Inc., as well as those from traditional asset managers and Wall Street banks, are grabbing an increased share of assets from their human rivals.  
Firms including Bank of America Corp.’s Merrill Lynch and Morgan Stanley have been targeting younger investors and those with more moderate wealth, offering their robo services to those with just $5,000 in assets. Betterment has no minimum at all. By comparison, human advisors at Merrill typically take on accounts of at least $250,000 in investible assets, while the threshold at Morgan Stanley tends to be even higher.
 
Aite Group estimates the number of robo-advice clients will climb to 17 million by 2021 from 1.8 million in 2016, with a big chunk of those clients having less than $10,000 in investible assets.

As assets grow and technology advances, the expectation is that robo advisors will become more sophisticated. Future robos might have the ability to interview clients, instead of simply relying on answers from a risk-tolerance questionnaire to make fund recommendations. Some envision consumers accessing robo services though a platform that looks like Amazon.com, where the robo will be armed with troves of data to help it understand investors’ lives.

 
Today’s robo advisors aren’t really all that intelligent, says Tucker Balch, professor of interactive computing at Georgia Institute of Technology. They’re essentially simple programs doing what human advisors do, like making trades and rebalancing clients’ portfolios — just much faster and more frequently. Eventually, that will become the baseline expectation of every investor, he says, which will push firms to use AI to try to differentiate themselves from competitors. Portfolios managed through robots likely will stretch beyond index-tracking funds into more active approaches that aim to spot opportunities in, say, specific stocks.
 
Some robos are expanding their offerings already. Wealthfront in March added a higher-cost fund that uses derivatives to replicate a popular hedge-fund strategy known as risk-parity. Wealthfront and others also offer smart-beta funds, which weight stocks by factors other than traditional market capitalisation.
 
Human touch
 
Some say human involvement may always be part of the equation. Someone has to write the algorithms for robos, even if the algorithms then teach
themselves. And it is humans—the firms that employ them and the clients who hire them—who will chose which metrics to maximise.
 
Balch is among those who believe human designers can equip robos with ethics. As for a conscience? “The best AIs will ruthlessly strive to maximise their profit,” he says, with a conscience arising “as a consequence of what that formula looks like.”