The Chatbots’ Challenge to Education: Disruption or Destruction?

Abstract

The article addresses the positive and negative implications of the growing spread of chatbots based on large language models (LLMs) on instruction, learning, and assessment in education. It is based on extensive conversations with ChatGPT regarding pedagogy-related issues and relevant documents. Discussed are the challenges of chatbots like ChatGPT to educators—on the one hand, their potential to advance deep learning and the roles of the instructor and the school context in causing it to happen. On the other hand, it underscores the pedagogical drawbacks of improper usage of such chatbots and the instructional practices and school contexts that could escalate learning. Three school-culture components, namely classroom learning, teacher professional learning, and school leadership, are the essential aspects of pedagogical approaches that, in a particular constellation, could enhance and, in another, impede a chatbot’s potential to advance deep learning. The underlying theoretical framework is assessment-driven, contrasting assessment for learning (AfL) and assessment for grading, distinguishing assessment cultures from testing cultures. Patterns of chatbot usage that align with the principles of each culture are discussed. A sample of quotes from the conversations with ChatGPT is presented to support the insights gained from the chatting experience and the conclusions drawn.