Description:
While ChatGPT's linguistic capabilities have recently seen an explosion of interest in a variety of fields, its potential in finance, particularly investment, remains unexplored. The study examined ChatGPT's efficacy in selecting asset classes and evaluating diversification effects using a quantitative methodology. The results suggest that ChatGPT's selections are significantly better in terms of diversity index than randomly selected assets. Additionally, portfolios constructed using ChatGPT's selections outperformed those built on randomly selected assets. The study concludes that ChatGPT has potential as an assistant for portfolio managers, and contributes to a better understanding of the role of language models in finance