The author is the Founder and Executive Chairman of Dimensional Fund Advisors. Booth his business at the University of Chicago and his school was named after him.
Can artificial intelligence help stock pickers? More specifically, can investors use AI to determine the fair value of stocks and bonds? Considering it’s possible, many would say yes.
I think my AI is better than others. My AI is the market.
For example, select stocks. Please check the price. Why that exact price? This is because the same number of buyers and sellers think that buying and selling at exactly the right time will result in a good deal. They make decisions using all available information, both public and private. Markets are the world’s largest information processing machines and determine the price of all publicly traded stocks and bonds.
These prices are set in an environment where no one knows what will happen. In that sense, it is the greatest giant model in human history, inferring the performance of each company’s stocks and bonds, and constantly evolving.
Despite AI expectations, I prefer to accept market prices rather than prices from algorithms. Large-scale language models, the type of AI that powers tools like ChatGPT, are intended to understand and produce human-like text, not to predict future results.
While it can generate potential scenarios based on learned patterns, it is difficult to account for unknown factors and real-world changes that appear outside of the training data. As such, they are truly “artificial” whereas the market consists of real human intelligence and millions of decisions made by market participants.
Indeed, AI and algorithmic trading can help execute trades. But there is no reason to believe that AI will fundamentally influence how people think about stock prices in the near future.
The market is surprisingly complex. With so many other concurrent inputs, no one knows exactly how much any given piece of information will affect the price. However, the market ensures that prices are the most accurate representation of the current value of stocks and bonds. Anyone can use it for free. How wonderful is that?
This is not just my opinion. There is plenty of evidence to support it. In fact, this was his fifty-year-old theory, and it’s been proven over the years. Google’s “Efficient Market Hypothesis”. Even better, let ChatGPT explain.
Still can’t believe it? Now let me ask you another question. Do you think you can hire a manager to implement a strategy that uses AI to consistently pick stocks that outperform the market? After you pay, probably not. If they have a superior AI that actually predicts stock prices better than the market, why would they share that information with you?
What do you get? You can have a good experience without worrying about such things. Based on about a century’s worth of data, the stock market returns about 10% annually, which is 7% above inflation. It was before and after computers, before and after the Internet, and even before and after World War II. I think it’s natural that it will continue after AI. This is because our AI is a “collective intelligence” that includes and improves on artificial intelligence.
To be clear, I celebrate the innovation that this moment can represent. As I have seen many times over the past 50 years of my career, many players use the latest advances in technology to improve their companies and build new ones. By buying the market, you get a piece of every publicly traded company.
If you’re still not convinced, we asked ChatGPT, “Is it safer to trust market pricing mechanisms than relying on AI models to spot mispricings in stocks and bonds?”
Here’s the answer I got on the day I asked the question: “It is generally safer to rely on market pricing mechanisms than to rely on AI models to spot pricing errors in stocks and bonds. and incorporates all available information into asset prices, making it difficult for a single investor or AI model to identify mispricing and consistently outperform the market .”
So if you can’t trust me, trust the AI that tells you not to trust the AI in the market.
