The productivity debate around AI is focused on the wrong variable. Paul Krugman raised it this week: cognitive damage, not productivity gains, may turn out to be the more consequential effect of widespread AI adoption.

The concern is not speculative. We already have reasonable evidence that outsourcing navigation to GPS degrades spatial reasoning. Outsourcing calculation to spreadsheets changed how accountants think. Outsourcing first-draft writing to AI will change something about how people write and, by extension, how they think.

This is not a Luddite argument. I use AI tools extensively. The point is that these are not neutral efficiency gains. They are substitutions that reshape cognitive habits at scale, and the downstream effects on workforce capability, professional judgement, and institutional knowledge are genuinely unknown.

Organisations that are thinking clearly about AI adoption should be asking not just "what tasks can AI do?" but "what skills do we risk atrophying if we let it?" The distinction between capability and dependency is worth drawing clearly, before the data forces the point.

#AI #FutureOfWork #Leadership

https://lnkd.in/gh7gKV-Z