Saved in:
Bibliographic Details
Main Authors: Erdem, Orhan, Hassett, Kristi, Egriboyun, Feyzullah
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2411.07031
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929718180184064
author Erdem, Orhan
Hassett, Kristi
Egriboyun, Feyzullah
author_facet Erdem, Orhan
Hassett, Kristi
Egriboyun, Feyzullah
contents We evaluate the reliability of two chatbots, ChatGPT (4o and o1-preview versions), and Gemini Advanced, in providing references on financial literature and employing novel methodologies. Alongside the conventional binary approach commonly used in the literature, we developed a nonbinary approach and a recency measure to assess how hallucination rates vary with how recent a topic is. After analyzing 150 citations, ChatGPT-4o had a hallucination rate of 20.0% (95% CI, 13.6%-26.4%), while the o1-preview had a hallucination rate of 21.3% (95% CI, 14.8%-27.9%). In contrast, Gemini Advanced exhibited higher hallucination rates: 76.7% (95% CI, 69.9%-83.4%). While hallucination rates increased for more recent topics, this trend was not statistically significant for Gemini Advanced. These findings emphasize the importance of verifying chatbot-provided references, particularly in rapidly evolving fields.
format Preprint
id arxiv_https___arxiv_org_abs_2411_07031
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Evaluating the Accuracy of Chatbots in Financial Literature
Erdem, Orhan
Hassett, Kristi
Egriboyun, Feyzullah
Artificial Intelligence
Econometrics
We evaluate the reliability of two chatbots, ChatGPT (4o and o1-preview versions), and Gemini Advanced, in providing references on financial literature and employing novel methodologies. Alongside the conventional binary approach commonly used in the literature, we developed a nonbinary approach and a recency measure to assess how hallucination rates vary with how recent a topic is. After analyzing 150 citations, ChatGPT-4o had a hallucination rate of 20.0% (95% CI, 13.6%-26.4%), while the o1-preview had a hallucination rate of 21.3% (95% CI, 14.8%-27.9%). In contrast, Gemini Advanced exhibited higher hallucination rates: 76.7% (95% CI, 69.9%-83.4%). While hallucination rates increased for more recent topics, this trend was not statistically significant for Gemini Advanced. These findings emphasize the importance of verifying chatbot-provided references, particularly in rapidly evolving fields.
title Evaluating the Accuracy of Chatbots in Financial Literature
topic Artificial Intelligence
Econometrics
url https://arxiv.org/abs/2411.07031