For all the fab headlines about driverless cars, many Americans have only the most remote grasp of artificial intelligence on the move. Coming soon to a theater near you: synthetic actors, generated by scanners and…
For all the fab headlines about driverless cars, many Americans have only the most remote grasp of artificial intelligence on the move.
Coming soon to a theater near you: synthetic actors, generated by scanners and voice analysis. Already on the wall: a computer-rendered Rembrandt, so realistic that it has fooled art experts. Meanwhile score one for the gamers, as one machine trained itself to learn and master a brick-breaker video game by playing 10,000 practice matches in an hour.
But with investing by AI, the enthusiasm and awe are arguably far less intense, given that few have much experience with it — while the thought of entrusting one’s hard-earned pay to a handful of silicon chips conjures images of computer crashes and hell-bent hackers. Even if the introduction of AI into many aspects of our lives is inevitable, should investment be one of them?
Indeed, recent Wall Street history has shown that computers are sometimes too precise for their own good.
“It’s the lemmings-to-the-sea story all over again,” says Keith Baker, a CPA and mortgage banking faculty member at North Lake College in Irving, Texas. “If investors totally give over decision making to the machine and they all use the same data and similar algorithms, then they might all give a sell signal at the same time, depressing the markets and causing a permanent loss of capital.”
In fact, we’ve seen something like this happen more than once with automated trading. Baker points to the February episode where the Dow Jones industrial average plunged 700 points in 20 minutes. But there is a flip side: “Those who have the best artificial intelligence, combined with an evolved machine learning ability, will start winning more often in theory,” he says.
To that end, a company that has made many investors rich points the way to even more gains, at least if you follow its high-tech hypothesis. Gordon Moore, the co-founder of Intel Corp. (ticker: INTC), spelled it out in his famous theorem, now known as “Moore’s Law.” Decades ago, he hypothesized that the processing power of computers would double every two years.
How right Moore was — and when it comes to how AI will change investing even in the short term, it’s a literal case of let the chips fall where they will. Around the time of the first iPhone, computer chips smashed the 1 billion-transistor mark for the first time. Today, we’re at 8 billion — and it’s possible to fit 6 million transistors in the period at the end of this sentence.
“The speed of innovation in our space is incredible,” says Mike Kerins, founder and CEO of RobustWealth, a digital wealth management platform headquartered in Lambertville, New Jersey. “Improvements are made almost every day to make digital investment tools better, faster and more robust. It’s an exciting time to part of the innovation curve.”
“High-frequency traders, fund managers and other sophisticated investors have used software to interpret massive amounts of data and follow simple algorithmic rules for decades,” says Stephen Heitzmann, co-founder and CEO of Altruistic Investing, a startup in the financial technology sector, also known as fintech. “But it is not AI in the sense that it can take the data and adjust its own decision-making process as new information is received. This ability to adjust itself is true machine learning, or AI.”
Thus we’re still in the infancy stages of investment via AI, Heitzmann points out. As for what’s possible — and this is where it gets exciting or overwhelming, take your pick — true machine learning can move into the realm of making future forecasts. That’s something stock mavens have had crystal-ball dreams about for ages.
Think of how Amazon.com ( AMZN) collects data to pinpoint what you’re most likely to buy based on what you’ve bought before, a phenomenon known as predictive analytics. Sounds great, right? Well investors, we’re not there yet. And in some ways, we may never be because humans are frustratingly impossible to predict.
For example, even the smartest computer may not know how to contend with a dumb leadership team, a hostile takeover or other uniquely human irrationalities. It can’t take advantage of cryptocurrencies that soar without reason (unless you count, perhaps, a greed stampede). Nor can it calculate how Elon Musk’s bizarre behavior and cult of personality impact the share price of Tesla ( TSLA).
“Human advisors have a good sense of the value of their client’s time and when to make the call,” says Jason R. Escamilla, CEO and chief investment officer at ImpactAdvisor in San Francisco. “When tax reform passed last year, days before the year’s end, we were able to save our clients thousands of dollars each with tax-smart, last-minute action items. There was no iPhone app making those phone calls.”
“I expect that over time, human advisors who have greater experience and more sophisticated investment strategies will regain the upper hand,” says Tyler Hay, CEO of Evergreen Gavekal, an international, independent registered investment advisor serving high net-worth clients. “This year, we have seen a number of notable roboadvisors throw in the towel, such as Hedgeable and LearnVest.”
“At this stage, AI can provide a competitive advantage but we cannot purely rely on it without the constant monitoring of an ethical and human mind,” says Ruggero Gramatica, an AI specialist and founder and CEO at Yewno in Palo Alto, California.
Yet human vagaries can also play on the investor side of the equation, too. Says Gramatica: “On the one hand, one could argue that machines lack human judgment. On the other, we could say that machine learning systems can minimize human bias by allowing the construction of data-driven decision-making processes.”
So what’s the answer? Maybe it’s a one-two punch of man and machine: Computers supply data-driven insights, while seasoned professionals leverage them to the max based on years or decades of experience: an algorithm or real intelligence that’s so simple, you can calculate it on one hand.