Using GPT-4 as a guide during inquiry-based learning

• Steffen Steinert ^[1] ; Karina E. Avila ^[1] ; Jochen Kuhn ^[1] ; Stefan Küchemann ^[1]
  1. [1] Ludwig Maximilian University of Munich

     Ludwig Maximilian University of Munich

     Kreisfreie Stadt München, Alemania

     
• Localización: The Physics Teacher, ISSN 0031-921X, Vol. 62, Nº. 7, 2024, págs. 618-619
• Idioma: inglés
• Enlaces
  • Texto completo
• Resumen

  • Artificial intelligence (AI) systems like ChatGPT can be used in education to create physics problems1 and guide inquiry-based learning, a collaborative approach where learners actively construct knowledge2 through problem solving.3 This approach promotes deeper cognitive processing and enhances long-term retention, as opposed to the passive reception of information.2 We demonstrate this method by implementing a multipart task where students, guided by GPT-4, can self-evaluate and explain the results of applying a current to an iron wire. The responses given by ChatGPT stimulate the student’s cognitive and predictive abilities while maintaining the AI as a neutral facilitator, forming a crucial step in the predict–observe–explain cycle of inquiry-based learning.