Leveraging large language models for educator feedback: A ChatGPT-assisted approach to improving teaching practices in higher education

In the era of artificial intelligence (AI), higher education is undergoing profound pedagogical transformation. This study investigates the feasibility and effectiveness of employing large language models (LLMs), represented by ChatGPT, as an auxiliary tool to enhance teaching practices and improve student learning experiences. A ChatGPT-based feedback framework was developed to enable educators to promptly optimize instructional strategies, enhance classroom interaction, and refine knowledge delivery. The framework provides actionable improvement suggestions, adaptive teaching recommendations, and reflective guidance. A semester-long controlled experiment was conducted to compare the outcomes of educators utilizing ChatGPT-assisted feedback with those employing traditional instructional methods. Results demonstrate that ChatGPT feedback significantly improved teaching adaptability. During the initial intervention phase, the mean score of the experimental group was 74.00, surpassing the control group’s 71.35 with high statistical significance (p < 0.001). In the final phase, the experimental group maintained a significant advantage with a mean score of 86.50 compared to 84.63 (p = 0.015). Student satisfaction scores exhibited a consistent upward trend, while the technology acceptance model survey revealed mean ratings above 4.0 across perceived usefulness, perceived ease of use, and behavioral intention dimensions. These findings indicate that ChatGPT-assisted feedback substantially enhances teaching adaptability and student learning experiences, while also achieving strong acceptance among educators.