In this interesting TEDx talk, Charlie Gedeon, a UX designer and university instructor, makes a strong argument that AI’s biggest impact on education isn’t personalization or smarter tutoring. It’s that AI is exposing how broken the incentive structure of education already was.
If we’ve spent decades telling students that the A+ is all that counts, not the process, not the struggle, not the thinking, then of course they’ll hand the work to a machine the moment they can. He frames it as a structural failure that AI didn’t create but is now accelerating.
The most interesting part of the talk, for me, is Gedeon’s concept of productive resistance. He and his co-founder ran experiments in their UX studio testing different interaction models: what if the AI asked clarification questions before answering? What if it assigned preparatory tasks before giving a full response? Each version introduces a different level of friction between the user and the instant answer.
Gedeon acknowledges we haven’t found the right amount of resistance yet. Too much, and users abandon the tool for a simpler one. Too little, and cognitive offloading takes over. That tension is real, and it connects to something I covered through Bastani et al. (2025), who found that students using standard ChatGPT improved during practice but performed 17% worse when the AI was removed. A version designed as a tutor, giving hints and not answers, eliminated the negative effect entirely. That’s productive resistance in action, even if the researchers didn’t call it that.
AI and Cognitive Offloading: Between the Individual and the System
Gedeon closes with a fitness analogy: learn what AI is good for, use it to assist your thinking, and build a habit of verifying its outputs. Fair enough. But he also argues for systemic change: Finnish six-year-olds study misinformation, and he sees no reason North American schools can’t treat students with the same intellectual seriousness. He wants more regulation, not less, especially when companies like OpenAI, Google, and Anthropic release their most powerful models for free during finals season, putting unregulated tools in front of students at their most vulnerable.
His closing question lingers: “Who does AI really help when we end up depending on learning with it?” I’d add one thought. Fan et al. (2025) found that students using ChatGPT produced better essays but didn’t actually learn better, they skipped the metacognitive steps that build real understanding. The tool opened a door. But the student never walked through it.
Watch the full talk HERE.
References
- Bastani, H., Bastani, O., Sungu, A., Geb, H., Kabakcı, Ö., & Marimane, R. (2025). Generative AI without guardrails can harm learning: Evidence from high school mathematics. Proceedings of the National Academy of Sciences, 122(26), e2422633122. https://doi.org/10.1073/pnas.2422633122
- Fan, Y., Tang, L., Le, H., Shen, K., Tan, S., Zhao, Y., Shen, Y., Li, X., & Gašević, D. (2025). Beware of metacognitive laziness: Effects of generative artificial intelligence on learning motivation, processes, and performance. British Journal of Educational Technology, 56(2), 489–530. https://doi.org/10.1111/bjet.13544
- Gedeon, C. (2025, September). Is AI making us dumber? Maybe [Video]. TEDx Talks, TEDxSherbrooke Street West. YouTube. https://www.youtube.com/watch?v=PLACEHOLDER
- Gerlich, M. (2025). AI tools in society: Impacts on cognitive offloading and the future of critical thinking. Societies, 15(1), Article 6. https://doi.org/10.3390/soc15010006
- Shaw, S. D., & Nave, G. (2026). Thinking fast, slow, and artificial: How AI is reshaping human reasoning and the rise of cognitive surrender. Working paper, The Wharton School, University of Pennsylvania. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6097646