The metacognition crisis

Metacognition is the ability to think about your own thinking. It governs whether you actually understand something or merely feel like you do. It is how you catch your own errors, recognise when your reasoning has gone off track, and decide when to change approach. It is arguably the single most important cognitive capability a person can develop, because it determines the quality of everything else: every decision, every judgment, every act of learning.

Metacognition develops through friction. It requires the experience of being stuck, of struggling with a problem, of formulating your own thoughts before seeing someone else's answer. When AI removes that friction, when the answer arrives before you have wrestled with the question, the metacognitive muscle never engages. You get the output without the process. Over time, the ability to generate that process independently atrophies.

Research is now documenting this in real time. Studies show measurable declines in critical thinking scores among frequent AI users. Researchers describe a pattern of epistemic laziness: the tendency to accept AI-generated output without genuine evaluation. Not because people are intellectually lazy by nature, but because the cognitive environment no longer demands the effort. The monitoring, evaluation, and regulation functions that Flavell identified as the core of metacognition are simply not being activated.

Here is what makes this particularly dangerous: most people cannot see it happening to themselves. That is the nature of metacognitive decline. If your ability to monitor your own thinking is degrading, you lack the very faculty you would need to notice the degradation. You feel productive. You feel efficient. You are producing more output than ever. But the quality of the thinking behind that output is quietly hollowing out.

The industry's response has been to build more tools that prioritise efficiency. Faster answers. More automation. Less friction. This is not a solution. It is an acceleration of the problem.