
Corrections
1:50 - I said students who use AI every day scored about 80 points lower. It’s about 40 points lower (38 and 42 in the source).
5:47 - I said “almost 1,000 high schools in Turkey.” It should be almost 1,000 high school students.
👉 CSS Demystified: https://cssdemystified.com
👉 Newsletter: https://kevinpowell.co/newsletter
Canadian teens who use AI for schoolwork every week are about two years behind the ones who never use it, yet other studies show students with an AI tutor learning twice as much. So which is it? I dig into the research on AI and learning, and why the struggle is the part you can't skip.
✅ My video on the PHP vulnerability that I mentioned: https://www.youtube.com/watch?v=cBD1AgkM1VQ
🎓 Research
✅ PISA 2025 results (OECD): https://www.oecd.org/en/publications/2026/09/pisa-2025-results-volume-i_5265bfb1.html
✅ Canadian teens who use AI are two years behind (Adam Dubé, The Conversation): https://ca.news.yahoo.com/release-ai-chatbots-harmed-teens-151349745.html
✅ Measuring actual learning versus feeling of learning (Deslauriers et al., PNAS 2019): https://www.pnas.org/doi/10.1073/pnas.1821936116
✅ Generative AI without guardrails can harm learning (Bastani et al., PNAS 2025): https://knowledge.wharton.upenn.edu/article/without-guardrails-generative-ai-can-harm-education/
✅ How AI assistance impacts the formation of coding skills (Anthropic): https://www.anthropic.com/news/AI-assistance-coding-skills
✅ AI tutoring outperforms in-class active learning (Kestin et al., Harvard): https://serl.fas.harvard.edu/publication/ai-tutoring-outperforms-class-active-learning-rct-introducing-novel-research-based
✅ From chalkboards to chatbots (World Bank, Nigeria): https://blogs.worldbank.org/en/education/From-chalkboards-to-chatbots-Transforming-learning-in-Nigeria
✅ The impact of generative AI on critical thinking (Microsoft Research + Carnegie Mellon): https://www.microsoft.com/en-us/research/publication/the-impact-of-generative-ai-on-critical-thinking-self-reported-reductions-in-cognitive-effort-and-confidence-effects-from-a-survey-of-knowledge-workers/
⌚ Timestamps
00:00 - Two years behind
00:37 - The asterisk on the Canadian data
01:06 - The "learn faster with AI" pitch
01:26 - We learn through friction
02:14 - Feeling like you've learned vs actually learning
03:15 - What happens when students get ChatGPT
04:25 - Anthropic's study on developers
05:53 - Why it's so easy to fall into this
07:54 - So why did some students learn twice as much?
09:32 - "But AI writes the code anyway"
11:34 - Trusting AI and critical thinking
13:11 - How to learn with AI without skipping the learning
14:09 - CSS Demystified
Altri episodi di "General Musings with Kevin Powell"



Non perdere nemmeno un episodio di “General Musings with Kevin Powell”. Iscriviti all'app gratuita GetPodcast.








