Resumen de El uso de modelos de lenguaje de gran tamaño en el aprendizaje de Física Computacional
Serena Di Santo, Michalis Skotiniotis, Ana Paula Millán Vidal, Carlos Pérez Espigares, Juani Bermejo Vega
Available Large Language Models (LLM) can understand and generate human-like text and auto-complete code. This gives them the potential to enhance learning and teaching experiences. However, their use in education poses challenges in maintaining academic integrity and accurately evaluating students’ progress. This study investigates the influence of Artificial Intelligence (AI), and in particular LLM, on students’ learning experiences in a Computational Physics course during the second semester of the 2023/2024 academic year at the University of Granada, Spain. We present an anonymous questionnaire to the students to explore their experiences in AI-assisted learning and their perspective on the introduction of specific training on the use of AI in the course. Data analysis shows that i) almost all the students used in-browser AI assistance during the course; ii) the perceived efficiency of AI in problem-solving is medium-high, especially for learning coding concepts and specific syntax, however, the students report that the code proposed by the AI often contains errors; iii) the vast majority of students would recommend introducing specific teaching material on the use of AI. This study is conceived as part of a longer term project that aims to promote an improved use of AI in alignment with academic honesty and in support of the development of critical thinking, and to ensure that evaluations accurately reflect students’ abilities.
