#Artificial Intelligence
Stock investors hopeful for artificial intelligence to revolutionize stock picking may be disappointed, as current technology falls short of making intelligent choices. A review of stock-picking GPTs in OpenAI's marketplace reveals a fundamental flaw: they cannot determine if a stock is expensive or cheap, a basic concept in investing. ZDNET's evaluation showed dismal results when GPTs were asked to identify the most expensive tech stock, highlighting their inability to grasp basic financial metrics.
The failure of GPTs to comprehend fundamental questions about stock valuation has significant implications. Not only do they consistently provide incorrect answers, but they also reveal a shared misconception likely ingrained in their pre-training data. Despite access to external resources, these programs lack the refinement to understand patterns of speech in stock-picking domains, rendering them useless for investment decisions.
The inability of GPTs to handle simple questions about stock valuation underscores broader concerns about their limitations. While not hallucinating, GPTs exhibit ignorance by conflating absolute stock prices with notions of expensive or cheap. This ignorance reflects a gap in understanding the language of investors, highlighting the need for more sophisticated approaches to AI training that capture the nuances of financial discourse.