I'm a chatbot, ask me anything: using ChatGPT to improve learning experiences

Abstract

Artificial Intelligence (AI) offers substantial opportunities and challenges in higher education. Given the evolving technological landscape, educators must ensure that students acquire a skill set encompassing both AI and traditional academic skills to enable them to succeed in their studies and future careers. We tested two groups of students, who each watched a recorded lecture on an unfamiliar topic. The first group used ChatGPT to ask questions and clarify content during the lecture, while the second group used Google search for the same purpose. We assessed the impact of these tools on the students’ cognitive load (germane and extraneous) and measured active learning through the number of questions students asked. We also used a post-test quiz, covering the breadth of Bloom’s Taxonomy, to evaluate the efficacy of each method. We expected that students using ChatGPT would experience lower extraneous cognitive load, higher germane cognitive load, and would learn content more effectively. Qualitative results demonstrated a notable preference for chatbots over search engines, due to the ease of locating specific information and obtaining insightful responses. Our findings suggest the potential of AI as a transformative tool in education, helping to enhance and deepen learning, while ensuring students retain ownership of their critical and creative processes. Leveraging the potential of AI and large language models may also serve a broader purpose: by personalising learning experiences to match individual students’ language skills, experience levels, and needs, AI may bridge the gap between the tailored support students want and the practical constraints that educators face.

Keywords

• GenAI
• cognitive load
• skill building
• artificial intelligence
• AI