Financial literacy of ChatGPT: Evidence through financial news

Ya Ling Chiu; Xuechen Gao; Hung Chun Liu; Qiong Zhai

doi:10.1016/j.frl.2025.107088

Financial literacy of ChatGPT: Evidence through financial news

Ya Ling Chiu
, Xuechen Gao
, Hung Chun Liu
, Qiong Zhai

Black School of Business (Behrend)

Research output: Contribution to journal › Article › peer-review

2 Scopus citations

Abstract

Our study explores the potential of AI to interpret financial news articles and support the development of effective investment strategies. The well-known AI model, ChatGPT, was employed to analyze a set of financial news articles published in 2022 related to the Taiwan stock market and assign a buying score to each stock mentioned. The daily average buying scores provided by ChatGPT largely reflect the overall stock market movement. Stocks receiving higher buying scores tend to outperform those with lower scores. Portfolios composed of stocks with low buying scores significantly underperform the market, suggesting that negative news have a notable adverse effect on subsequent stock performance.

Original language	English (US)
Article number	107088
Journal	Finance Research Letters
Volume	78
DOIs	https://doi.org/10.1016/j.frl.2025.107088
State	Published - May 2025

All Science Journal Classification (ASJC) codes

Finance

Access to Document

10.1016/j.frl.2025.107088

Cite this

@article{bf8f01bcbf4245c499e3c1f669d4a9e4,

title = "Financial literacy of ChatGPT: Evidence through financial news",

abstract = "Our study explores the potential of AI to interpret financial news articles and support the development of effective investment strategies. The well-known AI model, ChatGPT, was employed to analyze a set of financial news articles published in 2022 related to the Taiwan stock market and assign a buying score to each stock mentioned. The daily average buying scores provided by ChatGPT largely reflect the overall stock market movement. Stocks receiving higher buying scores tend to outperform those with lower scores. Portfolios composed of stocks with low buying scores significantly underperform the market, suggesting that negative news have a notable adverse effect on subsequent stock performance.",

author = "Chiu, \{Ya Ling\} and Xuechen Gao and Liu, \{Hung Chun\} and Qiong Zhai",

note = "Publisher Copyright: {\textcopyright} 2025",

year = "2025",

month = may,

doi = "10.1016/j.frl.2025.107088",

language = "English (US)",

volume = "78",

journal = "Finance Research Letters",

issn = "1544-6123",

publisher = "Elsevier Limited",

}

TY - JOUR

T1 - Financial literacy of ChatGPT

T2 - Evidence through financial news

AU - Chiu, Ya Ling

AU - Gao, Xuechen

AU - Liu, Hung Chun

AU - Zhai, Qiong

PY - 2025/5

Y1 - 2025/5

N2 - Our study explores the potential of AI to interpret financial news articles and support the development of effective investment strategies. The well-known AI model, ChatGPT, was employed to analyze a set of financial news articles published in 2022 related to the Taiwan stock market and assign a buying score to each stock mentioned. The daily average buying scores provided by ChatGPT largely reflect the overall stock market movement. Stocks receiving higher buying scores tend to outperform those with lower scores. Portfolios composed of stocks with low buying scores significantly underperform the market, suggesting that negative news have a notable adverse effect on subsequent stock performance.

AB - Our study explores the potential of AI to interpret financial news articles and support the development of effective investment strategies. The well-known AI model, ChatGPT, was employed to analyze a set of financial news articles published in 2022 related to the Taiwan stock market and assign a buying score to each stock mentioned. The daily average buying scores provided by ChatGPT largely reflect the overall stock market movement. Stocks receiving higher buying scores tend to outperform those with lower scores. Portfolios composed of stocks with low buying scores significantly underperform the market, suggesting that negative news have a notable adverse effect on subsequent stock performance.

UR - https://www.scopus.com/pages/publications/86000745190

UR - https://www.scopus.com/pages/publications/86000745190#tab=citedBy

U2 - 10.1016/j.frl.2025.107088

DO - 10.1016/j.frl.2025.107088

M3 - Article

AN - SCOPUS:86000745190

SN - 1544-6123

VL - 78

JO - Finance Research Letters

JF - Finance Research Letters

M1 - 107088

ER -

Financial literacy of ChatGPT: Evidence through financial news

Abstract

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this