A recent study by Pomona Institute has questioned the performance of artificial intelligence in stock market investment. Researchers analyzed exchange-traded funds (ETFs) that rely on AI for investment decisions and found that their overall performance was not ideal, even worse than that of the S&P 500. This study has triggered people's rethinking of the application of AI in the financial field and also reveals the limitations of AI technology in practical applications. This article will analyze the research results in detail and explore the challenges faced by AI in investment decision-making.
Recently, a study conducted by economics professor Gary N. Smith at Pomona College and student Sam Wyatt has sparked deep thoughts on the performance of artificial intelligence in the stock market. Although AI hype has been driving up the stock market, the fact is that many exchange-traded funds (ETFs) that rely on AI stock selection have not achieved ideal results.
Smith and Wyatt mentioned in a Scientific American article that they analyzed all publicly relied on AI systems to make investment decisions since October 2017. The results show that most of these funds performed worse than the S&P 500 index, which represents the 500 largest companies in the U.S. stock market. Research shows that among the 43 funds that rely on AI, only 10 performed better than the S&P 500, which means that AI has serious problems in stock selection.
To give everyone a better understanding of the performance of these funds, Smith and Wyatt summarized it. Funds that rely on AI have an average annual return of 5% lower than the S&P 500's 12.4%. Funds that rely entirely on AI and have no human intervention performed even more miserably, with 11 funds lagging behind the S&P 500, and 6 of them lost money when the market generally improved. Overall, the average annual loss of these 11 fully AI-driven funds reached 1.8%.
Researchers point out that AI is unparalleled in terms of data relevance, but it does not understand the meaning behind this data. They mentioned: "The fatal weakness of AI systems is that although they can find statistical patterns, they cannot tell whether these patterns are reasonable or meaningless. Only when AI algorithms can understand the meaning of words and their relationship to the real world," he said. They become reliable in important decisions, including investment."
Key points:
Most exchange-traded funds that rely on AI are underperforming the S&P 500.
Funds that rely entirely on AI have an average annual loss of 1.8%, failing to make a profit when the stock market is generally improving.
Although AI can find data patterns, it has not yet understood the actual meaning behind the data.
In short, this research results show that although AI technology has advantages in data analysis, there are still great risks in relying solely on AI to make investment decisions in complex financial markets. Future research needs to explore how to better combine AI technology with human experience and judgment in order to achieve better results in the investment field.