[:en]AI will Not Cause Structural Unemployment[:]
[:en]Last week I blogged about the false arguments for Basic Income, one of which is automation. The media has been talking about automation and its consequences on the job market for the past few years. As with many memes, this is driven by an original set of books and thinkers who identified AI as an emerging technology that will whip out many jobs – both white collar and blue collar. The media then fed on itself, generating more articles and books on the subject. “Newspaper X wrote about it, why don’t we have a big piece on it?” – said some editor somewhere. The thing with this whole automation creating unemployment is that there is no evidence for it.
Automation and technology obviously eliminate jobs; however, the consequence of this is higher economic productivity and wait for it, wait for it,…. more consumption! We humans don’t stop once we have one car, a decent size house and an annual vacation (see 1950s); no, that would be far too reasonable. We need more stuff! We buy a bigger house, a second car, a cottage, and more vacations!
Humans love stuff. After tens of thousands of years of starving to death in caves and mud huts, physical possessions bring comfort, security and status. When we do manage to automate something we just move on to some other problem. Here are some well known automations we have accomplished:
– Farming (unskilled labour) – something like 95% of the population was in farming.
Horse rearing, training, selling, riding – A lot of horses
– Drafting blue prints (semi-skilled labour) – replaced by AutoCAD.
– Making electrochemical film, selling and using it (skilled labour) – Kodak had 350,000 employees at its peak.
– Moving from manual handling of shipping material to standardized containers (unskilled labour) – lots of people!
– Car manufacturing (semi-skilled labour)
The list of eliminated or transformed industries is long and continuously growing. Many of these industries were absolutely massive and represented significant parts of our economy prior to being obliterated by technology. Yes, the jobs disappeared, but new ones appeared. The transition is tough, but we have done alright so far. A basic rule of economics is that if an asset has a utility it will be used. Humans are an asset, they will thus be used to do something – the question is whether they will be used for high value work or low value work. Think – café barista vs. aeronautical engineer.
The second part of the argument about automation is that we will run out of jobs for the displaced workers. I find this preposterous. Anyone who thinks we will run out of jobs simply has no imagination. Here are some straightforward ways we could employ many people:
– More teachers & better teachers
– Rebuild our crumbling infrastructure (just this is likely enough to employ enough people to replace all AI displaced jobs)
– Build high quality housing and more affordable housing
– More caregivers for our aging population
More aid workers and people to help in crises (Yemen, Myanmar,…)
And that is without even inventing new jobs like social media experts, sustainability advisors, artisanal coffee makers, micro-brewers, or Artificial Intelligence consultants! Will technology have a major impact on our communities, of course! Will we have to adjust our political and economic system to compensate for some out of work white collar people and secretarial jobs, yes. But is the world going to end or change dramatically, no.
Like any deployment of technology, AI will be deployed progressively. We should see a progressive impact – not a sudden overnight one. If we really think that automation is an emerging force that will lead to mass unemployment, we would already be seeing signs of this and especially so in the countries that are the most advanced in AI deployment. The countries with the best technology in the world – Germany, US, China, … all have the lowest unemployment rates. It is therefore difficult to see how we can both be creating unemployment through AI and have the lowest unemployment rates.[:]Published on November 25, 2018