Since its inception, artificial intelligence has prompted a debate: Is AI meant to replace human thinking or is it actually IA—intelligence augmentation?

While AI engineers may debate that fact, many financial services firms that were initially dismissed as “robo-investors” have recently added human advisors. MIT economics professor Richard Bookstaber put it simply in his book “The End of Theory”: “No man is better than a machine, and no machine is better than a man with a machine.”

The synthesis approach is based on the realization that both machines and humans have shortcomings. Humans are too emotion-driven and susceptible cognitive biases, while machines lack an understanding of human thinking.

Human Shortcomings

One major human limitation is the tendency to be driven by emotion. Most notably, fear and greed have always loomed at the twin hazards for investors. Such emotions cloud our thinking and lead to cognitive biases like irrational escalation, a.k.a. The Sunken Cost Fallacy, which prompts us to stick with an investment because we’ve already put time and money into it. Prospect Theory also dictates that the pain of losing an equal amount of money trumps the joy of winning it.

Ralph Keeney, a Duke University professor emeritus, said that most people make bad decisions because they don’t realize that decision-making is a skill that can be improved. “If you just started playing tennis and never took any lessons, you’d be better, but if you took lessons and learned the components of it—how to serve, how to hit backhand—then you’d play much better.” Instead, most of us think that because we have made decisions all our lives, we’re good at it.

But many of our decisions are based on heuristics. As Daniel Kahneman explained in his 2011 book “Thinking, Fast and Slow,” when someone asks you, “How happy are you with your life right now?” your heuristic question is, “What is my mood right now?” If someone asks, “How much would you contribute to save an endangered species?” your heuristic question is, “How much emotion do I feel when I think about dying dolphins?” Kahneman also recalls an instance where a chief investment officer of a large financial firm said he invested tens of millions of dollars in Ford stock because he had recently been to an auto show and had been impressed. In that case, the investor’s heuristic question, “What do I think of Ford cars?” overrode the more salient question, “What do I think of Ford’s stock?”

Proof of the human deficiency in investing can be seen when you compare managed funds to index funds. Just one in 20 managed funds beat an index fund, according to Standard & Poor’s research.

Machine Shortcomings

While machines are better at making cold calculation, they have one major blind spot: They can’t understand or simulate human thinking. In particular, machines are poor at understanding what makes us tick as investors. That’s why Betterment, for one, added human advisors as an option earlier this year.

“It comes down to the fact that we’re ultimately dealing with humans and the way each of us view the world and what we need is different,” said  Dan Egan, director of behavioral finance and investment at Betterment. While some people prefer dealing with an automated system and “actually get upset if we try to deal with them as a human,” Egan said, others want some kind of confirmation from a human “to make sure they haven’t clicked the wrong button.”

That will likely change. The same rigor that is applied to making financial decisions is increasingly being used to personalize advice. But at the moment, humans are much better at understanding other humans. For instance, within 0.2 seconds, humans can detect confidence in another person’s voice and we can detect anger in the first word of a telephone call.

The Man-Machine Mind Meld

Proponents of IA point out that even our best computer systems do little more than synthesize human decision making. The PageRank algorithm that Larry Page created to improve search results, for instance, was based on surfacing the pages that had been consulted the most. In other words, as John Markoff points out in his book Machines of Loving Grace, those results “essentially mined human intelligence by using the crowd-sourced accumulation of human decisions about valuable information sources.”

Of course, as Keeney points out, most of those human decisions are flawed anyway. Yet somehow, the combination of imperfect human thinking and cold machine calculation provide something more valuable than either could provide on their own.

Investing in securities involves risk and there is always the potential of losing money when you invest in securities. Before investing, consider your investment objectives and Betterment’s charges and expenses. Investments: Not FDIC Insured. No Bank Guarantee. May Lose Value. (c) 2017 Betterment LLC