Abstract

Since the advent of Chat Generative Pre-trained Transformer (ChatGPT), individuals have found new ways to leverage its capabilities to assist them in their personal and professional lives, including financial investing. Being a large language model (LLM) and generative artificial intelligence (GAI) tool, ChatGPT uses unsupervised deep learning methods on vast datasets which it leverages to provide responses to prompts by users (Nah et al., 2023). In the financial investing domain, ChatGPT can assist with researching companies, creating code to build forecasting models, and providing investment advice. Using news headlines, ChatGPT has demonstrated superior predictive capabilities of stock returns over more established methods (Lopez-Lira and Tang, 2023). With ChatGPT’s potential to enhance one’s financial investing prowess, providing greater insights into its capabilities is warranted. Using ChatGPT for financial investing purposes poses challenges as well. For instance, ChatGPT can hallucinate and provide incorrect information or be biased due to the training data used (Nah et al., 2023). Suboptimal advice could also result if prompts are not properly engineered. Therefore, research is needed to identify such issues for its safe and productive use in financial investing. To expand ChatGPT’s effective use, it will be important to identify its capabilities as well as its limitations. Therefore, we propose the following two research questions: (1) What unique capabilities can ChatGPT provide to enhance financial investing? and (2) What are the challenges associated with ChatGPT’s use in the context of financial investing? To answer these questions, we are conducting a qualitative study of user-generated content collected from Reddit’s online financial investment community regarding ChatGPT. Using a Grounded Theory approach, we are analyzing the dataset with two independent coders and using open, axial, and selective coding to identify categories, subcategories and a thematic explanation (Strauss and Corbin, 1998). We are also utilizing Yin’s Principles of Data Collection to provide support for reliability and validity (Yin, 1994). From our research findings, we aim to provide future research directions from the rich content uncovered in this exploratory study. We also expect to provide practical implications by identifying specific capabilities and challenges of ChatGPT’s use in financial investing endeavors.

Download

Share

COinS