coLearn-AI is a research platform designed to explore how artificial intelligence can support—not replace—student learning. Rather than focusing on generating answers, the system emphasizes the process of thinking, guiding students as they work through structured, collaborative activities.
At its core, coLearn-AI captures what we call an epistemic trace: a detailed record of how students develop their ideas over time. This includes their initial responses, revisions, interactions with AI feedback, and the timing and structure of their work within a group. By preserving this process, the platform enables new forms of analysis that go beyond correctness to examine how learning actually unfolds.
The system also introduces adaptive AI guidance, providing feedback that evolves based on student behavior. Early in a problem, the AI may ask open-ended questions; as time progresses, it can offer more targeted suggestions. This allows us to study how different forms of guidance influence student understanding, collaboration, and persistence.
The Summer 2026 project focuses on building this platform into a research-grade experimental environment, enabling controlled studies of AI-supported learning. The goal is to generate meaningful data and insights into how students learn, how AI can best support that process, and how collaborative learning environments can be improved.
This work sits at the intersection of computer science, education, and AI, and is aimed at producing both practical tools for the classroom and publishable research on the future of AI-assisted learning.