For centuries, mechanisation in various forms has had an impact on working practices.
From Jethro Tull and his seed drill in the very early 18th Century, through Newcomen, Arkwright and Faraday to Henry Ford’s first automated car production line in the early 20th Century, man’s ingenuity has spawned innovations that have brought untold benefits to millions but at the same time have wrought dramatic changes to the way we work.
Historically, the impact on overall employment levels has not been dramatic, as people previously employed in newly obsolete, labour-intensive industries have retrained and/or taken jobs in emerging new industries.
This principle is well illustrated by the 2005 film Charlie and the Chocolate Factory in which Mr Bucket’s job putting the caps on tubes of toothpaste is made redundant by a machine, but he retrains to become the engineer who repairs and maintains said machine.
With the exception of the World War II years, the proportion of US workers in manufacturing has been in almost constant decline since 1939, as illustrated in Chart 1.
However, the fact that major economies such as the UK and US are even now at or close to full employment, with millions of people employed in sectors of the economy that have grown hugely or simply didn’t exist 10 or 20 years ago, would seem to support the idea of job replacement.
That said, there is a legitimate debate to be had about the relative quality of some of those jobs and the labour force participation rate.
The changes brought about by the Industrial Revolution were so profound that it might appear bold – bordering on reckless even – to claim that any subsequent ‘revolution’ could have even more far-reaching consequences. Yet we are, of course, living through exactly such a ‘revolution’ – the Information Revolution.
At the heart of this revolution is the second most abundant element on earth – silicon – and the discovery of its properties as a semi-conductor, combined with the insight of Gordon Moore, whose Law states that transistor density (which we can take as a proxy for computing power) doubles around every 18 months to two years.
History may not repeat itself exactly (or even as tragedy followed by farce) but broad patterns do
tend to recur. It is wise to be very careful (not least in investment) when claiming that ‘this time it’s different’.
Yet there are reasons to at least ponder whether the impact of the Information Revolution on jobs might be more profound than anything that has gone before and, particularly, whether the historic pattern in which old jobs are replaced by new ones (at least for humans) in new industries may be broken.
Three main reasons why now is different
There are three broad reasons to think that this time could be different. First, ubiquity. Historically, machines or robots have been designed to carry out specific tasks in specific industries, essentially on an ad hoc basis.
Machines that work on this principle are already with us in the form of the intelligent personal assistants that are built in to smartphones and other devices.If you want a machine to screw the caps on tubes of toothpaste you design a machine with the necessary mechanical attributes. Information technology, by contrast, uses the same basic principles (all those zeros and ones) across any number of applications.
Just think what you can do with a simple laptop computer and the Microsoft Office suite and compare it with the mechanical tools that would be required to carry out the same tasks. That means the scope of activities that are open to the application of labour-saving techniques is essentially limitless.
Second, the sheer pace of progress – back to Moore’s Law. The implications of computing power doubling every 18 months to two years are staggering.
Recall the fable of the wise man who was invited for a game of chess by his ruler and offered anything he wanted as a reward for a valuable service rendered. He asked for one grain of rice to be put on the first square of the chess board, two on the second, four on the third and so on.
Eventually the board contained (or would have done had there been enough space or rice) no fewer than 18,446,744,073,709,600,000 grains. Apparently that’s more than enough to stretch from earth to Alpha Centauri and back, even if it’s short-grain!
Now imagine that applied to computing power (admittedly Moore’s Law has only been effective for c.40 years, but the number of chessboards we’re piling our rice onto has also expanded massively) and you get the idea of how rapidly computing power is advancing.
And with all that computing power comes data. In fact, of all the data in the world, around 90 per cent was created in the last two years.
This means computers are capable of holding vast quantities of information – and of processing it quickly. For many years computers have been competing against humans – and winning, even at a high level – in various games.
However, these games are structured and played to defined rules. The number of permutations of choices available to players may be vast, but it is finite and easily analysable given sufficient computing power.
Feed the computer enough data and give it the power to crunch that data and it will come up with the right answers. That brings us to the third element of the Information Revolution that could alter the employment market significantly – computers can be taught to learn without being specifically programmed to perform a particular task.
Detailed discussion of artificial intelligence is beyond the scope of this article but its implications could be profound.
Dealing with deeper problems
The ability of computers to effectively think for themselves, combined with the vast well of data to which they have access, could take us beyond a world in which the jobs that are replaced by technology are those that consist of performing repetitive tasks or are highly structured.
Rather than just repeating the process of screwing a cap on a tube of toothpaste for example, or identifying where the pencil marks are on a multiple-choice exam answer paper, machines could also be capable of performing tasks that require them to deal with unstructured problems.
In fact, machines that work on this principle are already with us in the form of the intelligent personal assistants that are built in to smartphones and other devices. Apple’s Siri is a well-known example, while Microsoft has Cortana and Google has developed Google Now.
All are able to understand questions posed in natural language (i.e. not coded in computer language)
and to provide answers on the same basis.
This development suggests that large numbers of clerical or even professional tasks that are still thought of as being the preserve of educated humans could come under threat.
There are examples of computers performing tasks that were previously the preserve of humans such as writing newspaper sports reports (although it’s not exactly Neville Cardus) which illustrate the potential. The scale and pace of change could well be more than the development of any number of new industries can cope with. There is of course a great irony in all of this.
The purpose of automation is presumably to improve efficiency and promote economic growth. But workers are also consumers.
Therein lies the problem. If everything is done by machines we might all have more leisure time (as promised by science programmes of 40 years ago) but how will we afford to live? Taken to extremes, most people won’t work and ownership of the means of production, to use an ominous phrase, will be concentrated in the hands of a few.
Economic questions
But if no-one can afford to buy what they are making, how does that help? This raises all sorts of social and economic questions, such as the merits of the universal or unconditional basic income, or universal income guarantee.
Under such a scheme, which was first proposed as long ago as the 1960s, everybody would be paid a flat basic income (funded by an extremely narrow tax base) regardless of whether they worked.
In fact it would provide a deliberate incentive not to work, with only those wanting or needing extra income working for it and competition for the few jobs that did exist actively discouraged.
That would represent a very topsy-turvy world compared to the one we are used to, but as anyone who has been paying attention to the investment world in recent years will have observed, why should that be a surprise? Then again, perhaps it won’t be different this time.
Simon Lapthorne is senior research analyst for Investec Wealth & Investment
