Redesigning Tertiary Educational Evaluation with AI: A Task-Based Analysis of LIS Students’ Assessment on Written Tests and Utilizing ChatGPT at NSTU

Abstract

The AI tool ChatGPT has shaken the academic system and challenges conventional teaching-learning methods. Using a mixed method approach, the author adopted a task-based analysis to measure the Library and Information Science (LIS) students' academic performance and integrity during the Artificial Intelligence, especially the ChatGPT trend. A group of 32 LIS final-year students were purposively selected as they completed all the basic courses of their discipline. A one-day session was divided into two phases – a written exam and an assignment through ChatGPT – held at the Institute of Information Sciences (IIS), Noakhali Science and Technology University (NSTU). Both examinations were administered using a questionnaire including a series of 15 tasks. A faculty member of IIS was tasked with reviewing the gathered answer scripts voluntarily. Results showed that it is too difficult to rank students' positions based on ChatGPT-generated tasks, as the participant who placed in 20th position (scored 5 out of 30) among 32 in the written examination, placed in 2nd position (scored 25) in ChatGPT-generated assignment. It is encouraging that the similarity indices (by Turnitin) reached up to 89%, which is unacceptable under university regulations. However, none of these assignments included in-text citations and references, which is questionable from the standpoint of academic ethics.

Ähnliche Arbeiten