Does ChatGPT show gender bias in behavior detection?

Ji Wu, Yaokang Song and Doris Chenguang Wu ()
Additional contact information
Ji Wu: Sun Yat-sen University
Yaokang Song: Sun Yat-sen University
Doris Chenguang Wu: Sun Yat-sen University

Palgrave Communications, 2024, vol. 11, issue 1, 1-8

Abstract: Abstract Recent advancements in Large Language Models (LLMs) suggest imminent commercial applications of such AI systems which serve as gateways to interact with technology and the accumulated body of human knowledge. However, whether and how ChatGPT contains bias in behavior detection and thus generates ethical concerns in decision-making deserves explored. This study therefore aims to evaluate and compare the detection accuracy and gender bias between ChatGPT and such traditional ML methods as Naïve Bayes, SVM, Random Forest, and XGBoost. The experimental results demonstrate that despite of its lower detection accuracy, ChatGPT exhibits less gender bias. Furthermore, we observed that removing gender label features led to an overall reduction in bias for traditional ML methods, while for ChatGPT, the gender bias decreased when labels were provided. These findings shed new light on fairness enhancement and bias mitigation in management decisions when using ChatGPT for behavior detection.

Date: 2024
References: View complete reference list from CitEc
Citations:

Downloads: (external link)
http://link.springer.com/10.1057/s41599-024-04219-3 Abstract (text/html)
Access to full text is restricted to subscribers.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:pal:palcom:v:11:y:2024:i:1:d:10.1057_s41599-024-04219-3

Ordering information: This journal article can be ordered from
https://www.nature.com/palcomms/about

DOI: 10.1057/s41599-024-04219-3

Access Statistics for this article

More articles in Palgrave Communications from Palgrave Macmillan
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().