Everything from this workshop in one place: the platform, a demo you can try on your own, project information, and papers about coLearn-AI.
The live platform. Instructors author activities, manage classes and groups, and monitor groups as they work.
colearn-ai.com ↗ Open coLearn-AIA demo instance where you can explore coLearn-AI activities without setting up a class.
colearn-ai.com/demo/aied2026 ↗ Open the DemoAn overview of the coLearn-AI project: goals, design, and how the system works.
jimskon.github.io/coLearn-AI-Information ↗ Project InformationResearch on collaborative learning, computing education, and AI-mediated learning environments.
computing.kenyon.edu/skonjp ↗ Research SiteWhen LLMs can produce correct answers with little effort, students risk cognitive bypass: correct results without the reasoning those results are supposed to reflect. The paper argues for system-level design that makes reasoning a condition for progress. It presents coLearn-AI as an example of that approach, with role-based collaboration, turn-taking, AI limited to a scaffolding role, and an append-only interaction trace.
Introduces coLearn-AI, which combines guided-inquiry pedagogy with AI formative feedback for group work in introductory programming. Instructors author structured activities and manage groups, and they embed AI guidance directives in activity files to keep feedback aligned with their teaching goals. The platform includes synchronous workspaces, turn-taking, and analytics dashboards.
Describes an assessment mode for the platform. In this mode it delivers work to one learner at a time, turns off formative AI help, and runs submissions through an AI first-pass grade constrained by instructor rubrics, with each score explained by reference to the student's work. Instructors review the AI's grades, adjust them, and publish final scores through an auditable interface.
The activity used in the hands-on session, ready to import or adapt.
The complete activity markup language, with every tag and option.
The research questions, the live Observation view, metrics, exports, and privacy.
Professor of Mathematics and Computer Science, Kenyon College · Karolinska Institutet
skonjp@kenyon.edu. Questions, pilots, co-designed studies, and shared activities are all welcome.