We are witnessing a new wave of automation. Computers and robots are becoming both cheaper and more advanced year by year, and they take over more and more work that used to be done by humans. We look at four scenarios for what role human beings may play in the highly automated society of the future.
In 2013, Oxford Martin School of Economics caused a stir with a report which estimated that roughly half of US jobs faced replacement by computers and robots over the next 20 years. More precisely, it was estimated that 47 percent of American job types faced a great risk of being automated, while another 19 percent faced a medium risk of being automated. According to the analysis, only 33 percent face little or no risk of automation. Among the ‘safe’ jobs, the authors Carl Benedikt Frey and Michael Osborne count work in health, law, finance, media and education. However, if we look at the automation that already is happening in these fields, with e.g. IBM’s self-learning supercomputer Watson, it seems naïve to imagine that these fields will remain ‘protected’. Few job types in the listed fields are likely to be entirely replaced by computers and robots, but new technology will make it possible for a few to do things that today are the work of many. Hence, unless there is a need for handling far more tasks, even these fields will be subject to increased unemployment due to automation.
In a more recent report from 2015, Frey and Osborne take a closer look at the issue. A recurring point is that automation generally increases inequality in a society. Since 1980, productivity in developed economies, measured as GDP per hour worked, has increased an average of 1.7 percent a year, but work wages, including employer-paid benefits and insurances, have only increased 1.1 percent a year. The difference has gone to the people who own the machines or speculate in securities – a result that reflects economist Thomas Piketty’s conclusion that income from capital investments generally grows faster than income from work. The Gini coefficient (a measure of income inequality) in developed economies has since 1980 grown from less than 28 to more than 33, and this reflects that the very richest in our societies experience enormous income increases while the majority experiences stagnating or even declining real wages. Frey and Osborne think that the sweeping automation wave we will see in the coming decades will increase inequality further by particularly benefiting those who own the machines, directly or indirectly through capital investments.
Two central uncertainties
The question is if it has to be this way. The future isn’t set in stone. Social developments aren’t shaped solely by large, unstoppable trends, but also very much by the decisions we collectively make. Hence, we need to think in terms of scenarios – visions of different, possible futures – to get the full picture of the challenges and, not least, opportunities that the changes will bring.
In the years following the financial crisis, most developed countries have experienced major unemployment. In recent years, economic growth has started to resume, but employment hasn’t increased as much as it was hoped. The general interpretation is that the renewed growth is due to a combination of increased automation and rapidly growing tech companies that don’t need many employees. The three biggest companies in Silicon Valley were in 2014 estimated to have a combined market capitalisation in excess of a trillion dollars. Distributed among their only 137,000 employees, this was about USD 7.3 billion per employee.
There is no doubt that automation will make a lot of labour obsolete. What is uncertain is whether the economic growth generated by automation, with some delay, will create as many workplaces as are lost. This has actually been the way of things since the dawn of the industrial age, in spite of the massive automation that followed the invention of the steam engine. When farming was industrialised, new jobs were created in manufacturing, and as manufacturing has become automated, new jobs have cropped up in the service and knowledge industries. Economic growth gives people more money to spend, which means that they can pay for products they couldn’t afford before, or have others do tasks that they used to have to do on their own. Automation generates economic growth, and economic growth generates jobs.
However, several scientists believe that the situation is different this time. As mentioned, half of current jobs are expected to be automated over the next 20 years, and there is no reason to believe that the pace will decrease in the decades after that. This is generally a result of the development of artificial intelligence that makes computers and robots able to handle increasingly complex tasks. According to extrapolations of computer capacity, an inexpensive computer which you can get for USD 1,000 may already by 2023 have a capacity corresponding to the human brain. This capacity is expected to increase a hundredfold every decade after that if advances are made at the past and current pace. This means that a growing group of people will become unable to perform any kind of job – mental or physical – that it won’t make more sense having robots or computers do. I will argue that this in fact is a development that has been underway for a long time. There are people permanently outside the labour market today who would have had no problem getting a job fifty or a hundred years ago.
The first central uncertainty for the coming automation wave is hence: Will as many new jobs be created over time as disappear, or will automation take over more jobs than are created and hence lead to increased unemployment?
The other central uncertainty is the risk of increased economic polarisation that Frey and Osborne mention: Will the economic growth only benefit a small elite, while the majority experiences declining wages or is thrown into unemployment, or will the increased wealth come to benefit all? Increased polarisation as a result of automation is no law of nature; it is a result of the priorities and decisions of our governments (and hence our voters). A real political effort to combat inequality – something that the UN in 2015 set as one of the most important goals for the future – could mean that everybody, or at least most, will benefit from the global increase in wealth generated by automation. The question is if political commitment can be mustered globally for such an effort.
Four scenarios for human work in the future
These two uncertainties frame a range of outcomes in two dimensions. One dimension represents the uncertainty new jobs versus unemployment (to put it simply); the other the uncertainty economic polarisation versus economic equalisation (put equally simply). In order to illustrate the breadth of this range of outcomes as best possible, I outline four scenarios below which represent situations near the extremes of the two uncertainties. The scenarios are hence not predictions of the future, but rather illustrations of how differently the future may turn out as a result of the decisions we collectively make today. They can be compared to street signs at a crossroads: If you turn east, you come to Easthaven; if you turn north, you come to Northcastle, and so on. Which destination you find most desirable will likely depend on who you are and how well you think you will do in the different scenarios.
The four scenarios are based on the following combinations of uncertainties:
Once you have read the scenarios, you might consider which scenario you believe is most likely to come true. Then you might consider which scenario you would prefer to live in. If your favourite scenario isn’t the scenario you believe in the most – what are you going to do about it?
As automation gathers momentum, wages on the labour market will be put under increased pressure. Many skilled workers as well as well-educated academics see their jobs disappear and are forced to take whatever work they can find, no matter how humble. The few who become richer because of automation are willing to pay for all sorts of labour – as long as it doesn’t cost too much. The middle class is squeezed, and society has been split into an upper class of 0.1-1 percent of the population, a ‘middle’ class of 10-20 percent, with the rest making up a ‘precariat’. This ‘precariat’ is the poorest segment in society, living with constant uncertainty in their everyday lives. Job security is an unobtainable pipedream, and even when working full time, it is difficult for most people to make ends meet due to low wages.
The situation isn’t unlike the Middle Ages, when a small elite of feudal lords ruled over large populations of vassals and serfs, with a modest middle class of merchants, guilded craftsmen, and priests. All serve the ‘feudal lords’, but some may, due to fortunate circumstances, create decent lives for themselves. Social mobility is low, as neither the upper class nor the middle class is interested in giving the ‘riff-raff’ any opportunity to challenge their position.
The creative society
Certain types of work aren’t easily automated. This is particularly true for creative work in a broad sense – research, development, design, communication, coaching, management, strategy, entertainment and art – since computers can’t be programmed to be creative, let alone innovative. As more routine jobs in production, services and knowledge work become automated, labour is liberated for creative work, and it turns out that there actually are enough jobs in the creative industries. All people are born creative – just observe children at play – and even though many have unlearned their creativity in the performance-oriented industrial and knowledge societies, most are able to rediscover it.
It was said jokingly about the service society that everybody would make a living cutting each other’s hair, and in the creative society we could say that everybody makes a living entertaining, inspiring, inventing and designing for each other – we have machines to do all the dull, routine work. Material wealth is taken for granted, since automation means that products aren’t much more expensive than the raw materials they are made from. This means that the focus increasingly is shifted towards immaterial, emotional and cultural wealth. They who own the machines do not become particularly rich, since the work that machines can do in reality becomes worth less and less. What’s valuable is the special and unique, and that’s all the things that only people can do.
The automation of manufacturing has shown that a few people, with the help of advanced machines, can produce far more than a lot of people without such help. It turns out that the same is true for the creative industry. We have long seen this in the music industry, where a few superstars dominate the hit lists and the music shows in radio and TV. Why settle for second best when everybody can have the best? In the old days, a lot of people had to slave for years to make an animated movie; today a single person with the right software can do it alone. Architects and designers can handle far more projects in half the time by having computers do the laborious parts of the work. In the knowledge industry, advanced expert systems take over a lot of work that used to be handled by doctors, lawyers, economists, and other specialists. This has created a superstar economy, where the elite in all industries take all the customers without leaving any for the many under the top layer. The winner takes it all.
The result is that it becomes increasingly difficult for most people to find work, even though neither ability nor willingness is lacking. More and more end up on public benefits, which keep getting reduced as more and more need support. Because the unemployed with their almost negligible income stop being interesting as consumers, the elite basically only create products for themselves, apart from cheap fast food and mass-produced entertainment to keep the masses from rebelling. Panem et circenses. If the elite want people to do various things, no matter how demeaning, it is not difficult for them to find volunteers – if there is even the tiniest chance that it can lift those people out of poverty. Private reality shows have become a popular pastime among the elite.
The hobby age
Automation takes over more and more jobs, but since automation also creates economic growth, this is not seen as a problem. It doesn’t make sense to force people to work when the machines can do the work better and more cheaply. Hence, society has chosen to make sure that everybody has a good life, where machines provide not just the bare necessities, but basically all that a heart could desire. Some communities introduce basic income; in others, all sorts of services are made free for people that don’t have an income. Why should people suffer in a rich society?
Nobody has to work, but few choose to lean back and do nothing. Voluntary social work is common, and many throw themselves at hobby projects, creating art or inventing stuff. This makes society richer in social and cultural capital, and in this way, citizens earn the rights to their free benefits. A few still manage to create enormous value and thereby amass enormous fortunes, but it is ensured that everybody will benefit from the development. As Buckminster Fuller said already in the 1950s: A few can create breakthroughs that can support everybody else; hence we should do away with the notion that everybody must work in order to justify their right to exist.