Exploring the AI–Human Interface for Personalized Learning in a Chemical Context
Ted M. Clark, Nicolas Tafini
Abstract
This study investigates how students use ChatGPT to enhance their understanding of atomic symbolism, specifically the number of protons, neutrons, and electrons in an isotope. Semistructured interviews were conducted with 8 students as they independently used ChatGPT to answer questions before starting a General Chemistry course. None of the students had prior training in chatbot prompt engineering, nor had they previously used a chatbot for learning chemistry. The chatbot output was analyzed for accuracy and instructional relevance. A key difficulty for students before using the chatbot was understanding the mass number of the isotope. This study highlights the challenges posed by limited prompt engineering skills, which could lead to outputs with poor relevance. The importance of students’ prior knowledge and their reading skills were also crucial factors in effectively utilizing the chatbot responses. This work is novel in its focus on the AI–human interaction, emphasizing the importance of prompt engineering and the broader evaluation of how students use chatbots when integrating AI tools into their learning process.