VisibleAI : la plateforme qui redéfinit l'évaluation académique en offrant une visibilité sur l’interaction étudiant-IA

Mohsen Shahini's Kritik is presenting a compelling paradigm shift to higher education assessment: moving the focus from mere content recall ('Did the student get it right?') to assessing critical thinking and process fluency ('Is the student getting better at thinking?'). This is an evolution of academic evaluation, not just another tech wrapper. The core ingenuity lies in VisibleAI. Instead of participating in the traditional 'detection arms race'—where edtech platforms waste resources trying to flag AI-generated work—Kritik pivots toward radical transparency and integration. VisibleAI doesn't accuse; it *tracks*. It provides a granular, real-time view inside the student’s workspace, differentiating between text typed directly by the student, prompts submitted to various Large Language Models (LLMs), and content pulled directly from the AI chatbot itself. This technical approach is crucial because it reframes AI usage not as cheating, but as a quantifiable skill. Shahini's vision allows educators to treat AI use itself—the ability to craft effective prompts, integrate diverse sources, or synthesize machine output—as an explicit learning variable. When the process of writing becomes transparent, instructors can grade the *efficacy* of the thinking process, rather than just the final polished product. The platform’s expansion past detection into collaborative assessment via Kritik360 further strengthens its position. By enabling students to peer-grade and critique each other's use of AI, the system forces a human element back into the loop—a concept Shahini emphasizes throughout. The combined effect is a systemic shift: the tool doesn't replace the teacher; it enhances their observational capacity, allowing them to observe cognitive offloading in action and teach students how to *manage* AI dependencies responsibly. This model has significant resonance for the Canadian educational landscape. As universities increasingly grapple with the challenge of integrating advanced generative AI into curricula, a system that provides verifiable data on student-AI interaction is immensely valuable. It offers institutions a concrete mechanism—a way to move from panic and prohibition to structured pedagogical integration, helping students build mastery while acknowledging the technological shift.