Large Language Model Tools for Project-based Learning

Abstract

Project-Based Learning (PBL) has emerged as a prominent educational approach that im- merses students in meaningful, real-world tasks, fostering deep and lasting learning experiences. Unlike traditional instructional methods, PBL emphasizes a student-centered pedagogy, where learners actively construct knowledge through exploration, collaboration, and reflection. This approach not only nurtures a love of learning but also encourages students to form personal con- nections to their academic experiences, making education more relevant and impactful. How- ever, while PBL offers significant educational benefits, it also presents challenges for educators, including the complexities of designing and managing projects, assessing student learning, and balancing student autonomy with guided instruction.. The advent of artificial intelligence (AI), particularly large language models (LLMs), holds promise for addressing these challenges by en- hancing personalized learning, automating administrative tasks, and providing real-time feed- back. To ensure that these AI tools are sustainable and conducive to diverse classroom contexts, it is crucial to involve educators in the design process from the outset.

This thesis contributes to the intersection of PBL and generative AI by documenting a co- design process with interdisciplinary K-12 teachers aimed at integrating AI into PBL pedagogy. Through need-finding interviews, collaborative workshops, and iterative tool design, this re- search explores how AI can support teachers in implementing high quality PBL while maintaining the integrity of student-centered learning. We also investigate how this technology can augment the current roles of teachers without replacing them, and support their professional growth.

The thesis is structured around three key objectives: exploring the challenges educators face with PBL, co-designing AI tools that address these challenges, and proposing design guidelines for future AI tools in PBL classrooms. By refining the design of AI-powered PBL tools, enhancing teacher professional development resources, and ensuring these tools are accessible and equitable, educators will be better equipped to foster engaging, student-centered learning environments. These contributions not only encourage future research and development of AI educational tools, but also aim to foster a more immersive and constructionist learning approach in classrooms.