Enhancing Blended Learning through a Customized AI Agent and Evidence-Based Teaching Methods

<p dir="ltr">Introduction: This paper examined the use of a learning assistant chatbot (AI agent), built on Generative Pre-trained Transformers (GPT) with lecturer's developed contents, to enhance the student learning experience in a Digital Electronics module. Digital Electronics was delivered in a blended learning format, guided by an evidence-based teaching framework that uses validated knowledge from cognitive science and extensive research on effective and efficient teaching methods. The AI agent was employed as a virtual tutor to supplement student learning. The AI agent design was also guided by the evidence-based teaching framework. This involved providing a structured active learning approach in which students were provided with clear learning outcomes, instructional content that focused on the key subject concepts essential to building understanding, and focused question prompts to initiate their critical thinking. From this instructional framework, they were then encouraged and scaffolded to use their own initiative and critical thinking skills to experiment with this AI agent. The research aimed to analyse and evaluate students' experiences with this AI agent and identify ways to improve the learning process. </p><p dir="ltr">Methods: This study employed a mixed-methods approach to gather both quantitative and qualitative data. Data was collected from 146 first-year students across 8 classes in two Digital Electronics modules. The data collection process included distributing a questionnaire to all 146 students, conducting a focus group interview with a convenience sample of 10 students, and utilizing reflective practices from 3 teaching faculty members. Descriptive statistics were used to analyse and interpret the students' responses to the fixed response items. For the qualitative data, derived from the open-response items, focus group interviews and reflective practices, a broad thematic coding approach was employed. </p><p dir="ltr">Results: The research yielded positive results, particularly with the AI agent pre-trained on the module content. The results showed that the AI agent significantly enhanced students' understanding by providing rapid and accurate feedback. The AI agent was found to be effective and useful for various learning tasks, such as answering questions, offering explanations, and generating insights. Additionally, it can reduce teaching faculty's time in content preparation and delivery and enable a greater focus on developing students' capability for higher level thinking and other 21st century competencies. </p><p dir="ltr">Conclusion: The findings suggest through understanding students' perceptions, based on their learning experiences with the AI agent, teaching faculty can create instructional strategies to enhance both achievement outcomes as well as engagement and intrinsic motivation. By integrating AI technology, educators can achieve greater differentiation and personalization of instruction; an essential goal in modern education. In conclusion, a blended learning approach incorporating AI agent, when underpinned by evidence-based teaching practices, has the potential to further improve the student learning experience.</p>