Artificial intelligence (AI) and its potential to improve investing success has been a topic of much speculation. However, a recent study by researchers from the University of Florida has found that AI-powered trading models could generate returns exceeding 500%, which is a dramatic improvement. The researchers used ChatGPT, a highly advanced language model, to collect news on different companies forming part of the main US markets. They then analyzed whether each headline was good news, bad news, or neutral, backtesting these assessments and assuming investing one dollar in the next appropriate market period. ChatGPT proved to be more reliable than existing data-collection systems, mainly on account of its advanced language understanding capabilities.
The study found that while ChatGPT2 and ChatGPT3 were useless, the ability of the model to understand sufficiently only begins to show with ChatGPT4. The model worked well for long-only investment, with a return of 50%. However, its performance was significant when shorting stocks, leading to a return of 400%. Together, it returned 500%.
This is an exciting development as it could potentially revolutionize investment choices. Investment managers may overlook bad news more than they should, and ChatGPT’s nuanced approach to analyzing news headlines could change that. It is clear that sophisticated AI-driven models such as ChatGPT have the potential to improve investing decisions and deliver significant returns in a declining market.
The time period for the study is intriguing as it extends beyond September 2021, which is when ChatGPT4 stopped assessing data. However, the researchers’ model was crafted to assess news from October 2021 to December 2022, using a range of different companies forming part of the main US markets. The performance stands in stark contrast to buying and holding an S&P 500 ETF during the same timeframe, reflecting good news for investors willing to trust AI for their investment decisions.
The researchers caution that it is still early days for these models to be implemented in practice, but the possibilities are exciting. With the development of more sophisticated AI, such models could lead to significant and reliable returns. Ultimately, investors should consider the implications of such research when making investment decisions in the future.
