coLearn-AI

A platform for collaborative guided inquiry in which AI supports student reasoning instead of replacing it.

The problem

Generative AI makes answers cheap. A student can submit correct work while doing little of the thinking that the work was meant to develop. This cognitive bypass is a serious problem for computing education and for any discipline that assesses learning through what students produce.

The approach

coLearn-AI organizes students into small groups working through structured, POGIL-style activities. Its design rests on three ideas:

• Roles and turn-taking. Each group member has a defined role, so the work and the talk about it are shared.
• Constrained AI feedback. Instructors write the activity, the expected reasoning and the feedback guidance. The AI responds within those limits and prompts students to accept or revise their answers rather than handing them a solution.
• An epistemic trace. The platform records the sequence of responses, code, feedback and revisions, so instructors and researchers can see how a group arrived at its answer.

The platform has been used in introductory programming courses at Kenyon College and is being extended to other disciplines, including health professions education with colleagues at Karolinska Institutet.

Links

• Project informationOverview of the platform, its design principles and how to use it.
• CCSC Midwest 2026 workshopMaterials from the three-hour workshop Designing AI-Constrained Collaborative Activities for Computing Education Using coLearn-AI, including hands-on activities, authoring guides and resources.
• colearn-ai.comThe platform itself. A public demo is available.
• Source code on GitHubThe open-source implementation.

Publications

• Skon, J. P. (2026). AI-Mediated Collaborative Cognition: Designing Systems That Preserve Reasoning in the Age of LLMs. Journal of Computing Sciences in Colleges, 42(2), 17–25.
• Kim, B., & Skon, J. (2026). AI-Enhanced Guided-Inquiry Platform for Collaborative Learning in Introductory Programming. Journal of Computing Sciences in Colleges, 41(8), 206–219.
• Kim, B., & Skon, J. (2026). Automated Assessment using AI-Enhanced Guided-Inquiry Learning Platform. Proceedings of the 2026 ASEE Northeast Section Conference. doi:10.18260/1-2--58573
• Idowu, G., Deka, C., Skon, J., & Aydin, N. N. (2025). POGIL-ITS: An Intelligent Tutoring System for Guided Group Learning.