Assessing ChatG PT -Generated Comments for Video Content Viewing in Online Learning
Jiaqi Wang, Jian Chen, Qun Jin · 2024
This study proposes a framework for gen-erating individualized comments (i-Comments) in on-line learning utilizing the capabilities of ChatG PT to address the challenges of limited interaction and learner isolation in online environments. Our previ-ous studies have shown that i-Comments can reduce isolation, improve understanding, and increase sat-isfaction with online learning. However, the manual generation of i-Comments is labor-intensive, error-prone, and often includes subjective information. To overcome these limitations, we propose a framework for generating i-Comments by ChatGPT. It consists of five components: extracting textual content from videos using Whisper, entering learner's information, defining key i-Comments attributes such as timing, content, and quantity, generating contextually relevant comments employing ChatGPT, and then these com-ments are integrated into the videos. This approach can be expected to reduce manual labor, decrease human operational errors, and support multilingual content creation, demonstrating its potential for widespread application in various online learning environments. We conduct an experiment to validate the effectiveness of using ChatGPT to generate i-Comments.