Recently, a study on the application of AI models in the financial field has attracted attention. Research has found that mainstream AI models, including GPT, have unsatisfactory accuracy when processing U.S. Securities and Exchange Commission (SEC) filings and often fail to give correct answers. This directly points out the severe challenges faced by AI models in the application of strictly regulated financial industries, and also sounds the alarm for the practical application of AI technology.
Research from a startup found that GPT and other AI models often failed to answer questions correctly when analyzing SEC filings. The performance of AI models in applications in regulated industries such as finance must be higher to be practical. In addition, the non-determinism and uncertainty of AI models are also one of the application challenges.
The research results show that the application of AI models in the financial field is still in its early stages, and its accuracy and reliability need to be further improved to meet actual needs. In the future, improving the accuracy and interpretability of AI models and solving their non-deterministic problems will be key research and development directions. Only in this way can AI function safely and reliably in more fields.