Several active learning pedagogies require students to define learning objectives in response to a formulated problem, case or project. This should turn them into motivated, self-directed learners. In practice, however, the process of generating and shaping meaningful learning objectives is marred with challenges. In this project, we explore whether and how the use of Large Language Models can act as a scaffold supporting students in their search for interesting and meaningful learning goals. Through literature review and experimentation CO-PILOT aims to identify a series of interventions with LLMs in class. We empirically test their effectiveness across 9 courses.
