I’m a fan of David Miller’s popular weekly blog, “Diary of a Fund Manager”. Miller is Investment Director at Quilter Cheviot Investment Management and always has interesting things to say about the markets and investment trends.
He recently posted a commentary about artificial intelligence—a warmup of sorts for his upcoming appearance at the inaugural NextGen Alpha Artificial Intelligence investor conference in Frankfurt, Germany. (Disclosure: I’ll be there, too, presenting on October 26th about harvesting sentiment using Artificial Intelligence.) In his opening paragraph he asks the question many of us in the financial world wonder about, too:
“What impact will artificial intelligence (AI) have on financial markets and how can I, as a conventional investment manager, take advantage of the accelerating rate of change we are surrounded by?”
AI is central to the BUZZ Index strategy, so I thought I’d take the opportunity to respond to David Miller’s post and try to answer his questions.
Dear Mr. Miller
I enjoyed your diary post on the topic of artificial intelligence. It raises many salient points about the transformation underway in how financial markets operate, and asks a number of questions about AI which I hope to answer for you.
A bit of background: I am the CEO of BUZZ Indexes, the first company to license its IP for use in an accessible exchange traded fund (ETF). The BUZZ US Sentiment Leaders ETF (NYSE/ARCA: BUZ) launched April 19, 2016 and tracks the BUZZ NextGen AI US Sentiment Leaders Index. BUZZ’s data science team deploys the latest frameworks of artificial intelligence and machine learning applications to mine investment insights from vast online content. The Index selects 75 stocks with the most bullish scores and rebalances monthly to provide exposure to large cap U.S. equities that exhibit the highest degree of positive investor sentiment. (So far in 2017, our Index has outperformed the Morningstar US Equity Fund Large Blend Category by 2.5%)
You posit that the question right now is “whether AI is better than humans in the game of investment, which has seven billion moving pieces (all of us) and also has to cope with the weather.”
In fact, no human can process all the information that helps put those seven billion puzzle pieces together. There can be no doubt about AI’s superior ability to source insights from the huge amounts of data we’re coping with in the markets and make better, faster decisions—without emotion coming into play.
To answer another of your questions: yes, absolutely—machines alone are good enough to win without human intervention. One example is the IBM computer Watson beating two human contestants on TV’s Jeopardy quiz show, in 2011. (Prior to its win, Watson spent four years “studying” Wikipedia and absorbing its massive amounts of data.)
Computers are also masters of chess and the ancient board game Go (which is so complicated and nuanced it’s been described as alien intelligence). So when you ask if the fund management industry needs AI, my reply is another emphatic “YES!”
Investing is about data, analysis, and decision-making. These three elements are exactly what AI is at its core. The days of the “star” portfolio manager are numbered when you consider that AI is now beating humans at poker. Why wouldn’t it make better investment decisions, too?
When you say that AI investment strategies are too “two dimensional” and therefore “of limited use in a three dimensional world”, I have to disagree. Winning at high-stakes poker proves that AI can make good decisions with imperfect information. Its ability is even better than humans—in crisis or normal environments—thanks to its vast information processing capability and objectivity.
The goal of AI is to train machines using the concept of neural nets (modeled on the human brain) to make decisions. Like human brains, AI learns by being exposed to new experiences.
Take driving a vehicle as an example. We learn the rules of the road and pass the test to get our license. But what actually makes us a good driver is having many real-world experiences, where we make decisions when confronted with something totally new. Like driving in a foreign country.
Now imagine if, instead of 25 years of driving experience (like I have) we can have the equivalent of one million years’ worth of driving “experience”, like an AI system would have. The AI system’s vast experience is compiled from all the self-driving cars, Ubers, and other vehicle devices feeding data “in”. Our ego and emotions might compel us to say humans with 25 years’ experience are still better drivers, but the plain reality is that the AI system would make far better (and safer) decisions.
You ask about the power of AI to alter markets and create more or less volatility. You wonder if it will change the cost of capital for the companies we invest in—those enterprises that drive the global economy.
The answer here is yes, but in ways we cannot foresee. Hedge funds are a good example. Thirty years ago we wondered how these “super smart” investors would affect the markets, and while there have been impacts (at times more volatility, at times less), overall, the market evolved to them. Capital and investment vehicles, innovation in product design, liquidity and opportunity all evolved to co-exist with the now thousands of hedge funds around the world.
I agree that the rate of change is accelerating and we need to keep pace as best we can. And agreed, attempts to model certainty are always doomed to failure.
But AI isn’t modeling certainty—it’s being trained to make decisions based on experience, just like us. Only it is learning with exponentially more “experience” (data) than we are ever capable of consuming!
One question I couldn’t answer is when you ask if AI is “better than other technologies designed to replace human intervention in the decision making process.” I have to admit I don’t know of any other technologies better positioned than AI that do this.
I look forward to meeting at the NextGen Alpha Artificial Intelligence investor conference in a few weeks’ time. I have no doubt that we will have a spirited discussion about the use of artificial intelligence, machine learning and alternative datasets within the investment management arena in the years ahead. .
Founder, BUZZ Indexes