With the endless stream of media promoting ChatGPT, we can assume most university students are using ChatGPT in some way on their own, or with instructor guidance, as they study for classes, write papers, and complete assignments. Plagiarism has always been an issue and if students are forbidden from using Chat GPT to generate essays, they still may be tempted to use it to brainstorm, write and improve their essays. They may then submit this work as their own without mentioning it was modified or written by Chat GPT. Detecting plagiarism has always been an issue for university faculty, to ensure the credibility of the institution, and with the advent of ChatGPT, faculty have encountered a new problem; detecting if students are submitting work they have actually written. Thus, this study attempted to discover whether faculty are able to detect ChatGPT generated paragraphs. To that end, one group of students used ChatGPT in any way they chose to compose their essays and another group wrote their essays without using ChatGPT. The essays were then collected and random samples were distributed to 50 faculty members. By using a questionnaire we then attempted to discover if faculty were able to determine if they could distinguish between human-peer review end products compared to Chat GPT end products. Faculty were able to use AI detection software to help them and also used their own intuition. This presentation will discuss the results and how they can be used to inform teaching pedagogy.

Authors:
Laurence Craven, American University of Sharjah, United Arab Emirates
Daniel Fredrick, American University of Sharjah, United Arab Emirates

About the Presenter(s)
Dr Laurence Craven is a University Assistant Professor/Lecturer at American University of Sharjah in United Arab Emirates

See this presentation on the full schedule – Saturday Schedule

Bookmark