Adam Riley and I wrote earlier this article about the cognitive cost of AI fluency. I want to add a layer to that argument.
The Alvesson and Spicer concept of “functional stupidity” – where organisations suppress challenge and substantive reasoning while still appearing effective – describes a problem that pre-dates AI.
What AI does is accelerate it.
When a first draft arrives already polished, the temptation is to move straight to refinement. Critical review becomes surface improvement. The first draft becomes the final position with better formatting.
This is how judgement erodes – not through stupidity in the ordinary sense, but through what Kant called organised immaturity. People stop being asked to think, so they stop expecting to.
In our consulting work, this shows up clearly. The teams getting most from AI aren’t the ones using it fastest. They are the ones who have deliberately designed friction back into the moments where judgement matters
most.
The cognitive muscle is built through wrestling with uncertainty. Not through reviewing what someone – or something – else has already wrestled with.
What would your organisation lose if no one had to draft anything from scratch again?
