I am following the election debates for the European Parliament and time and again there are ideas that labor productivity should be boosted and that is the only way to grow the economy.

But how relevant is the notion of "labour productivity" in the era of automation?

Of course, every automation tool is used by some human being (at least today). E.g. ChatGPT that writes computer code is used by human being and its result is translated into the output of a human being. Algorithm synthesis tool or automated math tool is producing highly sophisticated algorithms, but at some point the human being starts up this chain and uses its output.

So - human being is point in the chain. Some may argue, that human being is not essential in the chain (e.g. in the automated law or automated medical-diagnosis-and-prescription tool), but there can still be societal pressure that human being should be kept in the chain. Safety arguments are usually used for that, but more likely such argumentation arises from fear to loose job.

So - "labour productivity" is aggregate measure and may be indirect measure of growth.

But if we are trying to do the analysis, then boosting "labor productivity" usually is translated into education and skills boosting policies for the human beings - i.e. - into policies about human beings. But - I argue - maybe the biggest share of the growth comes from the productivity of the underlying tool and the human labor input may be constant or even diminishing? E.g., if the human being is kept by the law and obligations in the automated reasoning chains of the law and medicine, then all the productivity comes from the tool and the role of human being is superficial and may even be economically ungrounded.

So - what I am trying to say - the focus on the "labor productivity" and consequential ideas about human skill building is becoming less and less relevant in the automated economy. Instead - one should focus on building the better tools, better reasoning agents, AI.

Of course, labour productivity can be kept as the proxy measure but it may be outdated and too indirect and with little explanation value, but - as I showed - one should focus on the "tool productivity" and boosting the tool sophistication and performance.

This is my thinking and my question is - is my reasoning and invitation to refocus from the "labour productivity" to the "tool productivity" a sound one? Is there relevant literature about this? And can my reasoning explain the productivity paradox and refocus growth policies?

Added. One can argue, that someone should build tools and hence the human labour productivity is decisive. But there are 2 counter-points to this:

  1. There are nascent self-development and self-improving tools. Robots that explores their capabilities and simulates and learns skills. Large language models that improve themselves by self-prompts and re-evaluating their answers. Math synthesis tools. Automated chemical and biological laboratories, etc. There are formal models of such self-improving machines - e.g. Goedel machine.

  2. The layoffs of the past 1-2 years in the biggest technology companies shows that less and less people are employed in the tool development and - hence - more and more people are pressed to be tool users. This point invites heavy pressure to rethink the value-creation shares of the tools and human beings.


1 Answer 1


We've been in the "era of automation" for centuries now (witness the systematic technological underemployment called the "40-hour work week"), and yet labor remains the major factor in the cost of pretty much everything. New technology leads to reorganization; the reorganization proceeds until labor is once again the bottleneck. Productivity of labor is never not relevant.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.