Automation – Will it Destroy or Liberate Us?16th January 2017
In 1967, the American poet Richard Brautigan dreamt of a ‘cybernetic meadow where mammals and computers live together in mutually programming harmony’. The notion that technology would one day relieve us of our labours so that we might commune with nature was a product of the technological utopian movement which continues to ignite the imaginations of the world’s tech-obsessed idealists. Before Brautigan, the economist John Maynard Keynes hilariously predicted in 1930 that by now productivity would have matched all of our material desires, allowing us to work just 15 hours a week. Even before Keynes, everyone’s favourite Marmite philosopher Karl Marx had the prescience to theorise a concept he called the ‘general intellect’ which harnessed the combined intellectual endeavours of humankind to produce machines which would ultimately do away with human labour and bring about the advent of socialism.
It seems today, however, that the optimism which tinted these seemingly naive fantasies has vanished. Mainstream discourse surrounding automation tends to focus on the hollowing out of the labour force, or how ‘robots are stealing’ our jobs, as one recent Daily Mail headline put it. Faced with such an alarming and uncertain future, is it too late to rescue the dreams of a technological utopia?
A cursory examinations of the stats does not fill one with optimism. In the United States between 1997 and 2005, manufacturing output increased by 60% while 3.9 million manufacturing jobs were eliminated in roughly the same period. The reality is that American workers are now competing with robots as much as they are with Chinese sweatshops. Take textiles as an example. Fifty years ago one worker could operate five machines with each running a thread through the loom 100 times a minute. Now one machine can run six times faster with one human supervisor overseeing 100 looms, equalling a 120-fold increase in productivity.
It goes without saying that automation is not just an American phenomenon, nor is it bound only to affect the manual labourers of the world. Moravec’s paradox holds that artificial intelligence can solve high-level reasoning problems much more efficiently than human beings but basic tasks involving perception and mobility such as picking up objects and recognising faces can often be completed better by a one-year old. Therefore it does not necessarily follow that as automation gets more and more advanced, low-skill practical work will be the first disappear.
It makes more sense, in fact, to assume that the stock analysts and engineers will find their heads on the block before the plumbers and receptionists. Those who raise the issue of technological unemployment are not merely regurgitating the tired and debunked protestations of the nineteenth century Luddites. This time it’s different. Never before have the high skilled workers been so threatened.
What is clear is that the latest developments in science and technology are increasingly challenging our assumptions about the limits of AI and the uniqueness of human beings. It turns out even some of the most esoteric activities which we have always thought were the preserve of humankind’s eclectic consciousness can be outdone by a machine. In 2014, a computer named Iamus composed a classical music piece entitled ‘Hello World’ which was indistinguishable from vanilla, human-composed music, even to professional musicologists. Worryingly for me, Associated Press already uses computer algorithms to write many of its articles.
What is clear is that the latest developments in science and technology are increasingly challenging our assumptions about the limits of AI and the uniqueness of human beings.
Many of these algorithms are far from refined, to be sure. Just look up Sunspring, a short scifi film written by a computer, but could just as easily have been created by an intoxicated toddler. But technology has advanced by leaps and bounds which were unimaginable even a few years ago so there should be no reason to complacently dismiss the possibility that one day Oscars, Emmys and BAFTA’s could be collected by producers on behalf of AI screenwriters.
Despite the rapid progress of science and technology, a robot takeover still seems farfetched. After all, technology may destroy some jobs but it will create new ones we can’t even imagine yet, or so the argument goes. Recent studies have demonstrated that all across Europe, middle-income jobs are disappearing whilst the number of high-paying professions and low-paying service jobs have expanded. It is not necessarily mass unemployment we need to fear, but mass underemployment as the labour market becomes polarised between the elite professions and the low-wage, low-skill, insecure work for which it is cheaper to employ inexpensive labour than to automate.
The commercial introduction of self-driving cars, which is bound to happen in the next few years, will probably be the first readily apparent indicator of this sea change but the transport sector is just the surface. It is estimated by the Oxford Martin School that 47% of all jobs are ‘at risk’ of automation in the next 20 years. The BBC has even set up a page that will tell you how likely it is that a robot will be doing your job in the next couple of decades. Basically, if you want to work in a future-proof industry, join the clergy.
One thinker who dares to challenge the dreary consensus is the American economist Jeremy Rifkin who has written optimistically of a ‘zero marginal cost society’ in which automation ‘frees human beings from the market economy to pursue nonmaterial shared interests on the Collaborative Commons’. Supplementing this idea is the increasingly popular ‘universal basic income’ which in theory would undermine the principle of wage labour, allowing people to pursue productive activities without having to worry about where the next payslip is going to come from. One day this may prove to be the most efficient means of combating economic inequality if automation becomes as ubiquitous as some experts think it might.
However, there is still a problem. It may be tempting to embrace a world in which we can spend all day luxuriating on couches like Roman senators whilst mechanical arms feed us grapes but the truth is that for many of us work provides a sense of purpose and belonging. Technological utopians often envisage robots relieving us of the obligation to work so that we can all become poets and artists but not all of us want to become poets and artists.
The most testing problem which automation poses may not be the economic inequality but the social and cultural dislocation which could arise in a world in which, for the great mass of the human race, there is nothing meaningful to do.
Crucial to understanding the disaffection which has possessed so many blue collar workers in once industrialised regions across the West is that their old jobs in the factories, mines and shipyards provided a sense of fulfillment which cannot simply be replaced by retraining in IT or service skills. The most testing problem which automation poses may not be the economic inequality but the social and cultural dislocation which could arise in a world in which, for the great mass of the human race, there is nothing meaningful to do.
Of course, human beings are capable of creating their own meaning and certainly the most doom laden predictions underestimate the evolving imagination of humankind. Equally, we should not be flippant about the risks involved in automating away human beings, content in the assumption that because the cotton mill didn’t cause social collapse, no harm could befall us now. It is only by considering the dangers we face that we can determine whether in one hundred years time we will all be living under machines of loving grace or fending off the robot uprising.