Borrowed expertise: Why AI’s productivity boom may not survive the generation that built it
Borrowed expertise: Why AI’s productivity boom may not survive the generation that built it
Publish Date: 2026-07-10 09:07:00
Source Domain: www.brookings.edu
Here is a summary of the key points from the article using an unordered list for clarity:
* The productivity gains from generative AI are real and substantial, contributing significantly to increases in output per worker, especially for novices and low-skilled workers.
* However, these gains are largely coming from experts who have decades of experience building expertise through deliberate practice and feedback before AI existed. They know how to best utilize AI tools.
* The conditions that produced this expert class are eroding, and the pipeline for training new senior experts is being attenuated.
* This raises questions about the long-term sustainability of the productivity gains from AI, as firms are increasingly reducing junior staff hiring while maintaining senior ranks.
* AI compresses the gap for routine information retrieval tasks but widens it for tasks requiring judgment, creativity and the development of new knowledge.
* The article argues that AI, by enabling “inference-cheap” conditions, may reduce the incentive for cognitive development and innovative thinking over the long term, even as productivity appears to rise in the short term.
* The article raises concerns about the lack of emphasis on fostering cognitive development and tackling complex, novel problems within AI deployment.
* This could impact the long-term production of new knowledge, paradigm shifts, and breakthrough innovations as AI continues to reshape knowledge work.
* The author identifies opportunities across education, training, policy and tool design to better build and replenish the expert workforce to sustain future productivity gains.
