AI will automate tasks, not skills

Michael Hicks, Professor of Economics at Ball State University, USA, claims productivity growth, whether through automation, plant design or better-skilled workers, doesn’t kill jobs – it eliminates tasks:

  • First, hard, dirty and dangerous ones – think agriculture and steel-making where output continues to grow in volume but now uses a small fraction of the labour once employed
  • Then routine non-cognitive tasks e.g. assembly line work or truck driving, brick-laying, retrieval or processing of information, moving or tracking the movement of goods
  • Then routine cognitive, like medical diagnostics, detecting cancer


To date, not all tasks have been eliminated in existing sectors but the share of work done by manual labour has been slashed profoundly:

  • Few now carry water, split wood for the stove or hand-loom our shirts – or spend hours balancing the books at the end of each day
  • It is also claimed that US labour’s share of output overall has been falling over the last 50 years – from around 66% to 58% – and AI/ robotics will simply accelerate this trend


Technological progress is also relentless in all sectors:

  • It displaces workers in existing sectors by displacing many of the tasks they do – for example, over the last 30 years, the production of US steel has risen by about 10% while employment there has fallen by about 60%
  • It first invaded goods sectors, then basic service sectors, and is now encroaching other service sectors positioned at the professional or more personal end of the services’ spectrum


In the process, and against expectations, overall employment has not fallen

New and more interesting jobs have been created in existing sectors, jobs which involve working with the new technology – at the same time, new sectors have kept on appearing/ growing to meet human needs which could not be met beforehand

Over time, therefore, we humans have always found other, usually better and more interesting, jobs to do

We all possess skills that are hard to automate i.e. non-routine tasks, showing empathy, integrating and analysing quantitative and qualitative data at the same time, learning new, non-routine tasks

The problem is few training establishments – schools, colleges, universities and business schools – offer anything much to improve these skills, just as they (currently) ignore the small matter of ‘productivity improvement’


Leave a Reply

Your e-mail address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.