Large Language Models, Experiential Learning, and AI Simulations in the Classroom: The HistoryLens Framework

Abstract

This paper introduces the HistoryLens framework, which leverages LLMs to create interactive, primary source-based textual “simulations” that involve students in the process of analyzing, comparing, and critiquing historical narratives. Three case studies from courses taught at UC Santa Cruz in 2023 illustrate the iterative development of HistoryLens, highlighting both its limitations and its potential to encourage student engagement, foster historical empathy, and develop both historical research skills and practical exposure to creative uses of contemporary AI systems. The paper also discusses the challenges and design considerations associated with implementing LLM-enabled educational games and simulations and surveys relevant literature in this emerging field.