Tuesday, 5 September 2017

Are Robots and AI going to cause mass unemployment?

Futurists, Utopians and Basic Income advocates seem keen to emphasise that we are entering a new economic era that will change everything. This is the age of the robots, where machines and machine intelligence will make most, if not all, human labour redundant.

This sounds great in many ways, but I think we shouldn’t get too far ahead of ourselves, for several reasons.
Will this guy be taking the jobs of table-tennis players?

Firstly this clearly is not in the process of happening at present – the UK has morepeople in employment than ever before (over 32 million – including 75.1% of those 16-65). Productivity is not increasing rapidly, which might imply major technological advances.

Secondly, there are clearly plenty of useful things for humans to do and so we don’t need to entertain this argument until robot slaves really are taking care of all our needs and our (reasonable) desires.

Revolution or gradual evolution?

I don’t doubt that technological chance will continue to alter workplaces and lead some jobs to change radically and even disappear entirely. We see this all the time: docker jobs lost to containerisation, telegrapher jobs lost to radio operators, radio operators lost to digital communication. The Luddites rioted against weaving machines. Change can be fast-paced and while it generally benefits society overall in the long-run by making us all richer, there are sometimes lost skills and in some cases, great hardships for affected individuals and communities.

I can certainly envisage computer technology, driven increasingly by machine learning, taking over more and more tasks currently done by humans. However, I don’t expect that all jobs will be taken overnight, or even for the foreseeable future.

A recent McKinsey report draws the conclusion that only about 5% can be completely automated in the short-to-medium term. Most workplaces will of course be revolutionised if the promised developments really do occur, with jobs changing focus along the way. However, this does not imply there will be mass unemployment.

Put simply, there are plenty of useful things that humans can do. New jobs spring up all the time, often relating to the new-technology. The idea of someone working in “social media” for example would make no sense to someone fifty years ago. More complicated machines also need more expert engineers to create, monitor and maintain them.

Other jobs can also be done by more workers to better effect. Many tasks that are done on a voluntary-only basis at present could become professions. People could spend more time caring for the disabled, for example, including spending time with them rather than taking care of their basic needs. A favourite example of mine is that class sizes in schools and Universities could be reduced by a half or a quarter by doubling or quadrupling the number of teachers.

Resource limits, supply and demand

A further constraint on the computerisation of everything is that machines require power and materials. If machines become more capable and ubiquitous then it may be that the earth’s resources struggle to keep up and the prices of machine inputs increase. On the other hand, if people start losing their jobs then they will look for work elsewhere.

This process will constantly tip the balance back towards employing humans as they get relatively cheaper while machines get more expensive. There may end up being jobs that could be done by machines that humans can do more cheaply, at least when consumer preferences (for example to interact with a human rather than a machine) are considered.

As mentioned above, there seems to be plenty of demand for labour in the UK and plenty of people who would be keen to come to the UK to work. I don't believe this is a UK phenomenon either--other countries with increasing populations seem to find work for their populace to do.

Is hourly averaging redundant?

Some may suggest that my work-based proposal for fairer taxation and benefit calculation are irrelevant because very soon there will be no work to do. I completely disagree that the change will happen so quickly that it would not be worth seriously considering the proposal.

However, to the extent that revolutionary labour-market change is coming, I would argue that my proposals is the best available way to manage this change in a fair manner.

If we see work as a good the availability of which is gradually being reduced then it would need to be rationed out during the period of reduction. This can be done very easily with hourly averaging by slowly reducing the maximum amount of hour credits that can be claimed per week/month. Those who work longer than the maximum will be taxed at a higher rate for those additional hours, which would then give them an incentive to work fewer hours, allowing others to pick up the slack. This shares out the work without banning people from working longer if it still suits them and their employer (though with the advantage that one or both parties will be paying a higher rate of tax or pay for the privilege).


To answer the question at the outset, I don't think Robots and AI going to cause mass unemployment any time soon. Technological advancements will change the labour market, as will changing tastes and environmental depletion.

However, I want to emphasise two things from this. The first is that we don't need a Universal Basic Income on this basis any time soon. Furthermore, instead of being a problem for my hourly averaging proposal, technological developments make it more attractive to manage such changes in a fair manner.

No comments:

Post a Comment