Artificial Intelligence, Investing, and Our Thoughts
I will start by saying that ChatGPT did not write this post. It probably could have, perhaps even better than I have done, and certainly in less time than it took me.
For those who have no idea what I am talking about, ChatGPT is the latest hot Artificial Intelligence (AI) tool, which is an online chatbot that can respond to input by users. You might compare this to Siri or Alexa, but with vastly more capabilities. ChatGPT can write essays, haikus, thank you letters, and messages of condolence. You can ask it to write a script for a scene in a movie where Jerry Seinfeld and Vin Diesel team up for Fast and Furious 12. You can provide it buggy computer code and ask for help correcting it. It’s only limited by our own imaginations, and safety measures embedded into the system overall.
How does it do this? In a way, we don’t know. With artificial intelligence tools, we don’t write the instructions for the computer to follow. Rather, we provide billions and billions of samples of text, from books to online forums, blog posts to scientific articles, and the computer itself learns and builds a recipe to cook up whatever dish we request. We understand the method for how the computer learns, more than the result of its learning.
AI is the hottest forum topic out there right now, and it is revolutionary. Scientific communities, economic forums, educational institutions, and dinner tables around the world are now trying to get their hands around the tools that we, as humans, have created. The release of ChatGPT for text generation, and to a slightly lesser extent, DALL-E for image creation, was the Kitty Hawk moment. The plane can fly and the world will never be the same.
There are important conversations to be had around ethical conundrums in virtually every domain with the onset of AI, but since Integris is an advisory firm, let’s focus on how this changes the investing landscape. This will happen in one of three major areas, AI in businesses, AI in investing, and AI in advising.
AI in Business
Much like the blockchain mania that spread during the start of 2021, there will soon be a litany of companies investing in AI with a limited understanding of its practical use for their business purposes. There is always truth at the heart of every narrative shift, which is why it is so tantalizing. The world is saying, “AI is important and businesses need it,” and it will be hard to be a business owner and not have a response ready when the question “so what are you doing around AI?” arises.
Not every company that utilizes AI will be successful. My father-in-law is an incredible artisan, and when I bought my first house, he gave me a collection of amazing tools. I quickly learned, however, that the right tools don’t guarantee success when they are placed in unskilled hands. The beliefs that “AI is remarkable” and that “not every company that uses AI will be successful” are not mutually exclusive. Think of the titans of the technology industry today, and remember that they were not all around at the debut of the internet. As you are bombarded by solicitations at dinner parties, in mailers, and in emails to invest in “the next big thing in AI,” know that it’s okay to say no. And if the fear of missing out starts gripping you, know that if you own a diversified portfolio of stocks and bonds, you will likely be increasing your investment exposure to AI gradually as thriving companies adopt the technology and new companies join a representative index like the S&P 500.
AI in Investing
The use of AI in investing strategies isn’t necessarily new. There have long been investment funds that operate with minimal human interface, trading based on algorithmic instructions. Algorithms are, in a way, the nascent stage of AI. To me, the distinction lies in the creation process. Humans write an algorithm, which is a decision making process that gives instructions to a computer. In AI, we take the next step: we give the computers data and allow computing processes to create decision-making structures and refine them over time.
One problem with creating an AI tool for investing is that there is a constraint on the amount of data that we can provide. Consider the billions and billions of pages of internet content used to train ChatGPT and DALL-E, and then consider that the totality of capital market data available, although significant, is much more limited. The New York Stock Exchange started with a signed agreement of 24 stockbrokers in 1792. 231 years of history, maybe 58,000 days of potential trading data. Even if we assume records are in pristine condition, the amount of data pales in comparison to our records of written word.
Another hurdle to overcome is answering the question of whether the theoretical returns from an AI generated investment decision process will truly outperform a diversified and passive investment strategy over time. That, in and of itself, cannot be answered until we have a historical frame of reference to compare strategies, which really needs to be decades long to be statistically significant.
AI in Advising
I wouldn’t be surprised if we see a company pop up called (AI)dvisors in the not-too-distant future, where you can speak to a digital avatar about your financial picture. If this current iteration of AI is the least capable version that will ever be, it’s not hard to believe that there will soon be meaningful competition in the advisory career space. The cutting edge is sharp and it’s finally starting to reach the softer skills required by wealth managers.
Much like the data problem in the investing space, there also exists a data problem for advising. Client situations are inherently unique, and while there are general themes of advice that reign true for many households, there is an emotional aspect to what advisors do. Sometimes, “sub-optimal” strategies mesh better with a client’s life priorities without undue risk to their financial plan. The human brain has thankfully evolved to help advisors perceive these circumstances in which diverting from the mathematical science of financial planning would be in our client’s best interests.
There will most certainly be AI tools that assist in the financial planning process, but I believe (and hope) that we human advisors are many years away from being pushed into obsolescence.
The material provided herein is for informational purposes only, and should not be construed as advice, including, but not limited to tax, legal, insurance, investment, or retirement advice. For your specific planning needs, please seek the advice of Integris Wealth Management, your tax accountant, attorney, insurance agent, or other professional as appropriate. Investing involves the risk of loss.