Andrew Y's Senior Project Blog

Project Title: Effectiveness of AI financial prediction algorithms in comparison to human investors BASIS Advisor: Ms. Ainslie Internship Location: Remote Onsite Mentor: N/A

Project Abstract

With the increased usage of generative artificial intelligence in all facets of life, and the increased interest among the average population in exploring the financial markets for possible financial freedom, it is only inevitable that these two would cross paths. The enticing possibility of simply asking an artificial intelligence algorithm which stock to buy and getting an answer in seconds is promising. However, there has been no research conducted on whether or not this method of predicting the market is actually effective. While there have been studies that examine the current use of artificial intelligence in finance, they do not compare the performance of A.I. to another prediction algorithm. That is what this study will attempt to answer. The performance of artificial intelligence, when it comes to predicting stock movement, will be compared to the performance of all human investors on the market. This will be done by gathering a consensus prediction for a certain amount of stocks and comparing them to the prediction artificial intelligence provides. These can then be compared with a statistical test, in order to determine which one is more accurate. It is nearly impossible that either algorithm will produce predictions that consistently align with actual stock movement. Therefore, simply having a smaller difference from the real performance is what will be considered accurate. Finally, this can help with the exploration of new avenues of artificial intelligence in finance and provide assistance with answering key questions about which specific prediction algorithms are more effective, and how those algorithms can be enhanced.