AI and the Future of Financial Planning: What You Need to Know
8 minutes read time.
Do you use AI in your daily life? You might be tapping into the technology without even realizing it. According to a recent survey, only 33% of people think they use AI services or devices when in reality, 77% actually do. From helping to plan out a vacation to accessing virtual assistants like Siri and Alexa, the tasks AI can assist with run the gamut. But what about using AI for financial planning?
Ron Shevlin, Chief Research Officer at Cornerstone Advisors and Senior Contributor at Forbes has been closely tracking the AI boom and what it means for money management. Shevlin says when we talk or think about AI as it relates to financial guidance, we’re talking about what are known as “machine learning tools.” Or, in other words, technology that will do an analysis of data and then produce guidance.
WATCH Your Money Map: AI and the Future of Financial Planning: What You Need to Know
From helping to plan out a vacation to virtual assistants like Siri and Alexa, the uses for AI are many. But what about using it for financial planning? On the next episode of “Your Money Map,” Jean Chatzky sits down with Ron Shevlin, Chief Research Officer at Cornerstone Advisors and Senior Contributor at Forbes, to talk about the advantages–and the pitfalls–of using AI as you map out your money goals.
A recent survey from Forbes found that 43% of people use AI to get answers to their financial questions. The big question — particularly knowing the propensity of AI to occasionally hallucinate — is should they be doing so? If you’re the type to fire up ChatGPT and ask about things like which stocks to buy or the best credit card to apply for, Shevlin says there are several caveats to be aware of:
AI can’t take stock of your spending and saving habits
“When it comes to things like stock market predictions and portfolio analysis, there’s a lot of potential for AI,” says Shevlin. In fact it’s really good at analyzing historical data, but as Shevlin explains, when you start asking it for future stock market and investment projections, things can go sideways. “If you’re saying, ‘What should I do,’ that’s when I think the tools really break down,” he says. “We don’t have the real-time [data] and we don’t have that feedback loop.”
Similarly, he notes that for the average person whose financial life is not dominated by their stock market investments, but dominated by their spending and saving habits, AI isn’t a great tool. Why? In part, he says, it’s because the technology doesn’t have the data flow or ability to really have an impact by analyzing the average American’s financial life.
Unless you’re really great with the prompts, AI doesn’t take your individual life into account
In other words, AI — because it can see volumes of data — may be pretty good at parsing big trends and giving generalized advice. But it doesn’t know you specifically. Imagine asking AI to put together an investment portfolio. You might be wondering how it would know about your risk tolerance, for example. Therein lies the problem. “They [AI tools] are not actually intelligent like you and I, where we hear a question and we interpret it to what we think you’re asking about,” explains Shevlin. “These models are basically just reading letters and words…not interpreting intention or meaning behind them.” The exception may be when there’s an expert asking the questions and giving the prompts. If you know enough to tell ChatGPT or some other large language engine the specifics of your credit history, how you like to earn rewards points, and other pertinent information and then ask it which card to sign up for, it has a better shot at success.
The data isn’t up to date
One of AI’s biggest limitations is the fact that the data used by the technology is only up to date to a certain point. “[Researchers and scientists are] continually updating the models based on the data,” explains Shevlin. “But [what you get when you ask ChatGPT or another form of AI a question is] not real-time access to the Internet. It’s only real-time access to the model dataset.” Imagine if, on a particular day, you asked AI whether or not a stock was a good buy. The technology wouldn’t have the ability to know that, for example, an hour before the company you’re considering buying was hit with a major lawsuit that could impact its price and future financial performance. It only knows the information up to the point its data was last refreshed.
AI doesn’t get high marks for accuracy
Here’s where those hallucinations come in. One of Shevlin’s favorite stories that illustrates the limitations of AI has to do with someone asking ChatGPT who holds the world record for crossing the English Channel on foot. “Sure enough, it came back with an answer,” says Shevlin. “It claimed that the guy who formerly held the record for swimming the English Channel held the record for doing it on foot.” When AI doesn’t know the answer, it is inclined to come back with something — so it may just make one up. You need to be discerning enough to know when that is happening.
AI is better at completing tasks than answering questions
Instead of using AI to get answers to financial questions, Shevlin says a better use of the technology is getting it to complete tasks. For example, you could ask AI to draft an email to your internet service provider, requesting that they provide you with a lower rate. ‘It’s not forward-thinking, but it’s how consumers’ financial lives will get better because of these tools,” says Shevlin.
The bottom line
What will the future hold for AI and financial management? Time will tell. “The capabilities are there,” notes Shevlin, who says it will probably be 5-10 years before the technology becomes truly usable and impactful. “Timeliness of the data, the propensity for these tools to make up things…these are just bugs that kind of have to be worked out.” AI or no AI, try these simple tools to get answers on whether your retirement plan and savings will last.