Scaffolding Critical Thinking with Generative AI

86% of university students use AI weekly. New research asks: does that use sharpen critical thinking, or quietly erode it? A 2026 peer-reviewed framework from researchers at the Spanish National Research Council and the University of Melbourne proposes eight design principles for integrating large language models into higher education without undermining the cognitive processes that deep learning depends on. Published in Computers and Education: Artificial Intelligence, this is one of the most practically grounded frameworks to emerge from the current wave of AI-in-education research. Topics covered: - Three documented risks of unstructured AI use: cognitive offloading, metacognitive disengagement, and epistemic narrowing - Six essential intellectual processes that instructional design must protect, including inferential reasoning, evaluative judgement, and epistemic integrity - Eight actionable design principles, from preserving cognitive friction to balancing AI-free and AI-mediated task phases - Two practical classroom scenarios showing how these principles apply to real course design - Implications for community college faculty, instructional designers, and curriculum developers navigating AI adoption Vendrell and Johnston synthesize cognitive psychology, educational theory, and AI ethics to address a problem growing on every campus: students who use AI tools heavily may be producing more polished outputs while developing weaker reasoning skills. Research cited in the paper shows that students who engaged with a task independently before consulting a large language model produced significantly stronger work. That single finding reframes the entire faculty conversation from "should we allow AI?" to "how do we sequence it?" For community college educators working with diverse learners and constrained resources, this framework offers a language and a structure for moving AI policy beyond blanket restrictions and toward deliberate, pedagogically grounded integration. The question is no longer whether LLMs belong in higher education. It is whether we are designing learning environments that require students to think with AI, rather than instead of thinking at all. If you found this analysis useful, please LIKE the video and consider subscribing for daily insights on AI in education. Have you encountered AI tools in your teaching or studies? Comment below. I'd love to hear your experiences or concerns! Based on Scaffolding Critical Thinking with Generative AI: https://www.sciencedirect.com/science/article/pii/S2666920X26000342 Podcast Version: https://open.spotify.com/show/202FbaDxmPvbnjD8lxq6TK?si=4HZtptYXQjGR0sTMyDJPrg Connect with me on LinkedIn: https://www.linkedin.com/in/rlethcoe/ Video Summary created with NotebookLM by Ronald Lethcoe