DIGITAL LIBRARY
USING GENERATIVE AI AND GPT CHATBOTS TO IMPROVE INSTRUCTION
Brigham Young University (UNITED STATES)
About this paper:
Appears in: INTED2024 Proceedings
Publication year: 2024
Pages: 5451-5459
ISBN: 978-84-09-59215-9
ISSN: 2340-1079
doi: 10.21125/inted.2024.1410
Conference name: 18th International Technology, Education and Development Conference
Dates: 4-6 March, 2024
Location: Valencia, Spain
Abstract:
This presentation reports results from an ongoing study investigating the use of Artificial Intelligence (AI) and Generative Pre-trained Transformers (GPTs) to provide interactive tutoring and assessment of student learning. Technology integration can be beneficial in education, facilitating the instructional process and assisting teachers.

This study explored two specific areas of potential value:
1) instructional feedback systems and
2) assessing student learning.

Using technology to support teachers in these two areas is important as teachers often do not have the time to provide one-on-one instruction, assess students individually, and provide each student with feedback on their learning with areas for improvement.

Technology, specifically generative AI chatbots, have the potential to assist teachers in these areas. Generative AI is a relatively new educational tool that needs to be tested and better understood. Preliminary findings suggest that a GPT can accomplish the task but needs to be guided with detailed instructions for best results. This research found that technological limitations and points of potential concern can be addressed with carefully constructed guiding scripts that provide the GPT with instruction that includes an instructional purpose, interaction rules, and a learning guide.

Using instructional scripts to focus the GPT, this research pre-trained a generative AI system chatbot to interact with a student regarding pre-assigned reading tasks made available through an open educational resource (OER). The GPT was instructed to answer questions about the content and then quiz the student on the topic to test their understanding of the content.

Results from this study were promising. Based on the premise that students who used the resource were interested in learning and preparing for a course assessment, the GPT effectively prepared students in an interactive, friendly manner that students found beneficial. An evaluation of the resource found that students were accepting of the resource and that the GPT could effectively provide specific feedback and accurately assess students' understanding of the content when provided with proper instruction.
Keywords:
Technology Integration, Generative AI, Educational Technology.