Last month we explored agency – how AI doesn’t take anything from you, you give it away. This episode goes deeper into the mechanism: what actually happens in your brain when you delegate cognitive tasks to AI?
In this episode:
- What cognitive offloading is and why AI is different from previous tools
- The “inverse skills bias” – why AI helps novices more than experts
- What we gain (speed, reduced load) vs. what we lose (memory formation, skill development)
- The “inflated knowledge” problem – mistaking AI’s knowledge for your own
- Digital dementia vs. technological reserve – two competing hypotheses
- Vending machine users vs. directors – the critical distinction
- Why friction is the mechanism of growth
- “Desirable Difficulties” and “Productive Failure” frameworks
- The key question: Am I trying to get something done, or get better at something?
Core principle: Cognitive offloading isn’t good or bad – it’s a trade-off. Performance now vs. capability later. You decide which tasks to offload and which to struggle through.
Research Referenced:
- Benge & Scullin (2025). A Meta-Analysis of Technology Use and Cognitive Aging. Nature Human Behaviour.
- Grinschgl, Papenmeier & Meyerhoff (2021). Consequences of cognitive offloading: Boosting performance but diminishing memory. Quarterly Journal of Experimental Psychology.
- Pyke, Lunau & Javadi (2025). Does difficulty moderate learning? Quarterly Journal of Experimental Psychology.
- Chen et al. (2025). Effects of generative AI on cognitive effort and task performance. Trials.
- Ding et al. (2025). Productive Failure in Cultivating Clinical Thinking. Advances in Medical Education and Practice.
- Danaher (2024). Generative AI and the future of equality norms. Cognition.
- Zhozhikashvili et al. (2022). Parietal Alpha Oscillations: Cognitive Load and Mental Toughness. Brain Sciences.
- Allen (2024). Desirable Difficulty—Make Learning Harder on Purpose. Clinical Orthopaedics and Related Research.
- Grinschgl & Neubauer (2022). Supporting Cognition With Modern Technology. Frontiers in Artificial Intelligence.
Additional Resources & Mentions:
- Special thanks to Kyle Shannon, host of the AI Learning Lab and founder of the AI Salon, for his concepts of “AI slop,” “chain of craft,” and the floor vs. ceiling distinction. Learn more at thesalon.ai
- Kyle Shannon interview: Becoming AI Ready: How to Creatively Secure Your Future – AI Explored podcast
- Quote from Cisco VP Anand Sampath about humans being “pushed up the stack” sourced from The Rundown AI newsletter
Referenced Episodes:
- Episode 1: AI Isn’t Coming For Anything – It’s Your Responsibility