Sam Sivarajan is a wealth management consultant and the author of two books on investing and decision-making.
We may not yet have the ever-helpful robot Rosie from The Jetsons or the deeply malevolent Skynet from the Terminator films, but Alexa and Siri are our daily companions. Artificial intelligence has burst out of sci-fi screens and into our everyday lives.
How real or fast is this technological shakeup? Some suggest AI, with its emotionless and bias-less programming, might be the revolution the investing industry needs. Is it? Possibly in some ways, but probably not in most ways. Why?
First, hype has always outpaced reality in innovation. For example, in October, 2016, Elon Musk, the chief executive officer of Tesla TSLA-Q, announced that the first fully autonomous Tesla would drive from Los Angeles to New York by the end of 2017. By 2022, that goalpost had moved to 2023 and many experts believe the reality is still years away.
This is not to argue that technological innovation has not made huge strides. The push for autonomous driving has resulted in key features including adaptive cruise control, lane-change assist, parking assist and so on. But robot taxis are not yet on the horizon.
Second, AI is not some bias-free utopia that will completely liberate us from the shackles of sexism, racism, short-termism or any other form of ism that humans suffer from.
When Apple issued its credit card, provided by Goldman Sachs, many complained that credit approvals and limits were apparently sexist. For instance, Steve Wozniak, the co-founder of Apple, shared that he and his wife had common assets and accounts and yet he was granted 10 times the credit limit she was. Despite the outcry, it wasn’t that the algorithm itself was sexist. It was simply that the data that fed the algorithm were based on historical credit decisions that may have been discriminatory. U.S. Senator Elizabeth Warren, a fierce advocate of consumer financial protection, argued at the time that, “We’re all beginning to understand better that algorithms are only as good as the data that gets packed into them.”
With more and better data, these algorithms will improve, but the underlying design and data biases will persist – we might just be less aware of them. Out of sight but not out of mind. That isn’t necessarily an improvement.
Third, there are some tasks that are easier to automate or delegate to AI. Generating text or art or e-mail scripts is a neat trick and may reduce time and effort for many daily tasks. It isn’t, however, going to write the next Pulitzer Prize-winning novel or the next Grammy-winning song or even the next company annual report. AI’s ability, today, is to largely extrapolate from a large data set of past examples. The ability to create new solutions to new problems does not yet exist.
What does this mean for investing?
Emotions or biases can easily sway the selection of stocks for a portfolio, for example. Analyzing and selecting stocks for a portfolio unemotionally is something that AI, to a large extent, is perfectly suited to do and do well. After all, analyzing stocks involves crunching lots of data, financial and otherwise, to determine attractiveness. Could AI do some of the tedious legwork?
AI can do the number-crunching, but remember, while AI will use the available data, it will not question the validity of the data, so we still need the analyst to check: Can this be right? Could AI make better strategic business decisions? For instance, $10,000 invested in Tesla at its initial public offering in 2010 would be worth almost $3-million today. Could AI have done better than portfolio managers of the time or rival auto manufacturers in betting on Tesla? It is hard to see the spark of genius or insight that AI currently possesses that could do that.
The average investor relies on advisers, fund managers or indexing algorithms to select securities. However, navigating market fluctuations and avoiding behavioural biases is their primary challenge. Data show that while the S&P 500 index has delivered a 7.5-per-cent annualized return over the past 20 years, the average investor has only earned 2.5 per cent, barely beating inflation.
This performance gap owes to behaviour-driven factors such as chasing returns in bull markets and panic-selling during bear markets. Behavioural finance research has many examples of overtrading because of overconfidence, return-chasing behaviour or herdlike behaviour. While AI may help identify behavioural biases, it cannot address them effectively.
Like a doctor pointing out the need for a healthier lifestyle, AI may help identify investor biases, but improving investor behaviour requires a human touch. AI cannot motivate someone to exercise or overcome short-term fears while keeping long-term goals in mind. A coach or adviser can provide that guidance.
AI will undoubtedly be a valuable tool in the investment industry, but it cannot provide the reassurance investors seek about their financial well-being. It cannot answer, in a compelling and empathetic way, the one and only question they care about: Am I going to be okay?
The notion of AI revolutionizing the industry any time soon, or solving the biggest investor problems, is greatly overhyped. That doesn’t mean AI can’t, won’t or shouldn’t play a much bigger role in the investment industry. It should and, used the right way, the individual investor will reap the benefits. However, AI is not a panacea and it will not solve the big behavioural problem facing investors.
There is some truth to the quote attributed to Pablo Picasso, “Computers are useless. They can only give you answers.”
Be smart with your money. Get the latest investing insights delivered right to your inbox three times a week, with the Globe Investor newsletter. Sign up today.