Too Good to Be True? An Empirical Study of ChatGPT Capabilities for Academic Writing and Implications for Academic Misconduct

Presenter Information

Peter André Busch, University of AgderFollow
Geir Inge Hausvik, University of AgderFollow

Paper Type

Complete

Description

This study evaluated the capabilities of ChatGPT, a conversational artificial intelligence (AI) agent, to generate quality answers to exam questions that are indistinguishable from human-generated text, normally requiring high-level cognitive abilities. This potential capacity is a major threat to academic integrity since ChatGPT may be used for academic misconduct. To determine its capabilities, we asked 15 faculty members to assess a total of ten answers (two each) to an actual IS exam question, of which students wrote five, and ChatGPT generated five. The study found that ChatGPT can generate answers of generally good quality that may pass as human-written text by examiners. The contributions of the study lay in providing empirical results in a novel research area, providing recommendations for teachers, and suggesting ways forward for research.

Paper Number

1829

Comments

SIG ODIS

Recommended Citation

Busch, Peter André and Hausvik, Geir Inge, "Too Good to Be True? An Empirical Study of ChatGPT Capabilities for Academic Writing and Implications for Academic Misconduct" (2023). AMCIS 2023 Proceedings. 21.
https://aisel.aisnet.org/amcis2023/sig_odis/sig_odis/21