Our worry about technology taking our jobs is age-old. Now, with the exponential development in Artificial Intelligence, computer power, new software and robotics, the discussion is hot once again. In this blog, we take a look at the future of work and automation, and think there is plenty to discuss.
In March 2017, Fortune magazine told their readers that “Robots Could Steal 40% of US Jobs by 2030”. Robots stealing jobs undoubtedly sounds like a worrying future prospect. The background is a PwC report, “The Impact of Automation on Jobs”, that suggested that robots and automation could take 38% of the jobs in the US, 30% in the UK, 35% in Germany, and 21% in Japan.
Why is the potential job-loss higher in the US? The report points out one central aspect; the US has a large financial retail sector, with financial workers in more routine jobs that is likely to be automated. And think about it, the Automatic Teller Machine (ATM) was one of the first “robots” to “steal” jobs from the traditional bank teller, an employee that deals directly with bank customers. According to the ATM Industry Association, there are now close to 3 million ATMs in the world. How many jobs have these machines taken from bank employees, without us even noticing? Another “robot” that then “stole” more bank jobs was the Internet bank that allowed customers to do even more routine jobs themselves, such as paying bills, transferring money and opening accounts, thus enabling bank employees to focus on higher value add jobs such as selling mortgages and managing mutual funds. Now, as asset management is also being automated and run by algorithms at super-low fees (since running an index fund is really just routine too), the financial jobs are going elsewhere again. And what about the old ATMs? New technology, such as mobile payments, are taking their jobs as we move towards a cashless society with no needs for robots handing out paper bank notes. New robots are taking jobs from older robots, not only people.
Humans have long been worried about “technology unemployment”, and history shows that old jobs have always been replaced by new ones. The enabler between the old and the new job market is education and our ability to develop new skills. Some argue that the current technology development is faster and more far-reaching, as machines can do ever more of human work, mostly thanks to AI and progress in robotics. It’s a significant difference for humankind between technology historically replacing manual tasks, for example in agriculture and factory assembly lines, to now replacing human intelligence.
Today there is more private capital than ever around to further fuel the digitalisation of business and society. The case for investments in technology is strong, as automation enable companies to reduce labour costs (usually the largest single cost in a company), thereby increasing productivity and profits, freeing up resources for more innovation that will enhance competitiveness, and ultimately drive return on capital. That’s why, in the struggle between capital and labour, capital will once again be the winner in the digital era. This is ironic, since many private equity firms manage money for pension funds, and the investors actually acts as catalysts for job-loss, by making investments in technology that may displace the jobs for the ultimate beneficiaries of the pension funds (i.e. employees and future pensioners), in order to create returns on their pension capital. But as private capital is also used to grow companies and create new jobs, the net effect is hopefully positive.
An opinion piece in the Wall Street Journal, “The Challenge of Our Disruptive Era”, calls our time the largest economic transformation in recorded history. At the core of the change is jobs. They are no longer permanent. “We are entering an era in which we’re going to have to create a society of lifelong learners. We’re going to have to create a culture in which people in their 40s and 50s, who see their industry disintermediated and their jobs evaporate, get retrained and have the will and the chutzpah and the tools and the social network to get another job.” says historian Ben Sasse.
Many of these evaporated jobs are displaced by automation, something that we tend to generalise as robots. So, let’s sort out a few perspectives. First, what is really a “robot”, the machine that is “stealing” our jobs? Second, where is automation happening? Given huge inequalities in the world’s economies, is automation really a global phenomenon? And does automation affect all jobs? And thirdly, what about the short and the long term effects? Should we be worried, longer term?
What is a robot? The robot has become the general term for any technology that displaces human jobs. For example, it’s a chatbot in retail guiding customers, as we have shown in an earlier blog (“The Meaning of Bots”), it’s applied Artificial Intelligence used, for example, in video or text analysis (as we explored AI in a ”The Current State of Artificial Intelligence”), or an industrial robot used in car manufacturing. A “robot” can be both software or hardware, and usually a combination.
Where is automation happening? Another distinction is that job-losses due to automation is likely to happen primarily in industrialised, developed economies like the US, the UK and Germany, and less likely in poorer, less developed and digitised countries. However, the poorer countries are still likely to be affected in other, indirect ways by automation. For example, most of the outsourcing in recent years happened because manufacturing moved from rich (expensive) countries to countries with lower labour costs in Asia or Eastern Europe. If robots can produce even cheaper in rich countries, the manufacturing might move back again, then affecting poorer countries. And as we will see later, different jobs are also unevenly changed by automation.
Third, automation of jobs must be viewed both in a short term and a long term perspective. Short term, there will be job losses at the hands of automation, but long term these gaps will likely be filled with new jobs that we could not have imagined. This is the general idea anyway. For example, today, a clerk operating a computer with a web based ticket-handling tool at a customer support center is considered a rather low-skilled job, but compared to a farmhand in the agricultural society (a job now mostly gone), the computer operator is a very advanced job, unimaginable to the farmer a 100 years ago. In the long term, we have not much to worry about then, according to lessons from the past. Question is how long societies can survive the pressure of the short term job loss squeeze, especially if the short term effects, say the coming 20 years, are more dramatic than ever before in history.
A recent post in TechCrunch, “Technology is Killing Jobs, and Only Technology Can Save Them” refers to several studies that show how automation is indeed displacing humans, and argues that the only way to fight the job losses is to train the talent that we have. It also argues that technology rarely automates major occupations altogether. Studies could only find one job that had disappeared completely – elevator operators. The first jobs to go are usually the ones that we don´t want anyway, called the “three Ds” – dull, dirty and dangerous jobs. However, by reducing these jobs, automation also dramatically reduces the opportunities available to unskilled workers. But technology is also creating jobs. For example, collaborative robots is a major growth potential for the industry, leaving the machines to handle the “three Ds”, while human coworkers handle tasks better managed by people. An article in Packaging Digest, “What Are Collaborative Robots and Why Should You Care?” argues that the robot is just like a new member on the team.
These new team members are, however, not like you and me. For example, robots do not pay any tax (at least for now). Probably the most famous promoter of a robot tax is the world’s richest man, Microsoft’s Bill Gates. In a recent article, “The Robot That Takes Your Job Should Pay Taxes”, he explains how there will and must be taxes that relate to automation. “Right now, the human worker who does, say, $50,000 worth of work in a factory, that income is taxed and you get income tax, social security tax, all those things. If a robot comes in to do the same thing, you’d think that we’d tax the robot at a similar level.”, says Bill Gates. He thinks the tax could be used to temporarily slow down the spread of automation and use the tax to finance education and fund other types of employment for the humans displaced by robots.
A New York Times article, “The Long-Term Jobs Killer Is Not China. It’s Automation”, finds that around 2 million jobs were lost to outsourcing in China in the 2000s, but automation have reduced even more jobs. Compare this to the job-loss in just one of many US industrial sectors, steel production, where over 75% of the work force, or 400,000 jobs, has been lost because of new technology since 1962. President Trump can maybe force companies to keep production onshore, but it will be difficult to stop them from staying competitive by automation and innovation, and thus the jobs will disappear anyway. A study from Ball State University attributed only 13% of manufacturing job losses to outsourcing and trade while the rest to automation that enhanced productivity.
Technology and automation will destroy current jobs, and create new ones. To be able to take on these new jobs, workers need to keep up with technology change by lifelong education. To this end, both governments and companies have to rethink the way education is designed, defined and delivered. The Economist dedicated a recent issue to the pressing importance of lifelong education, and in a leader titled ”Equipping People to stay Ahead of Technological Change” they stated:
“When education fails to keep pace with technology, the result is inequality. Without the skills to stay useful as innovations arrive, workers suffer—and if enough of them fall behind, society starts to fall apart. That fundamental insight seized reformers in the Industrial Revolution, heralding state-funded universal schooling. Later, automation in factories and offices called forth a surge in college graduates. The combination of education and innovation, spread over decades, led to a remarkable flowering of prosperity. Today robotics and artificial intelligence call for another education revolution. This time, however, working lives are so lengthy and so fast-changing that simply cramming more schooling in at the start is not enough. People must also be able to acquire new skills throughout their careers.”
We assume the effects of robots and automation will be dramatic, but while there are many jobs and tasks that have the technical potential to be automated, that potential varies much across different types of work. Automation is thus not only unevenly distributed geographically in the world, but also looks different in various industries. Several studies are devoted to finding out which jobs are the most likely to to be automated in the short term. Such studies tend to get attention since they indicate 1) who is likely to be unemployed going forward (and needs to re-think and re-train), and 2) what are attractive future jobs (and thus influence career and education decisions). For example, in a previous blog we showed how Swedish university students are increasingly interested in AI-courses, since that field is perceived by young engineers as future-proof.
A study from McKinsey, “Where Machines Could Replace Humans, and Where They Can’t”, looks at the technical feasibility to automate jobs, as the percentage of specific activities in that job that can be automated by adapting today’s technology (not whatever technology available tomorrow), thereby identifying the most and the least “automatable” activities. What are typical automatable activities? McKinsey finds, not surprisingly, that the most automatable jobs are predictable physical work like an assembly line, food processing and packaging, generic “3Ds” jobs (dull, dirty and dangerous). Here, the technical feasibility of automation is estimated to 78%, suggesting that so many jobs or activities are likely to be displaced by “robots”. The hardest activities to automate are those involving managing and educating people (estimated to only 9 percent automation potential) or expert jobs involving decision making, planning, or creative work (18 percent). The McKinsey report indicates that mostly typical blue-collar jobs will be automated, while white-collar workers are less likely to be replaced by robots. This is not really the case, as other studies indicate.
In a much cited study from Oxford University, “The Future of Employment: How Susceptible are Jobs to Computerisation?”, the authors estimated the probability of computerisation for 702 occupations in the US, and found that about 47% of the total US labour force is at risk.
What about the Nordics? A report from the Swedish Foundation for Strategic Research, takes a similar approach as the Oxford study: looking at a large set of jobs in society, breaking them down into specific activities and assessing what activities can be automated by software. This is, in essence, the general digitalisation that goes on in society. The many parts in human activity that are gradually digitised. As we discussed in our blog post “Digitalisation and the Nordic Opportunity”, every enterprise software, e-commerce site and online payment platform together make up the digitalisation of both the society and its jobs.
The Swedish study analysed 109 different professions and found that 53% of today’s jobs (or 2,5 million of the total 4,7 million people employed today) are likely to be affected by digitisation in the coming 20 years. Professions that require creativity, social skills and persuasion are less likely to be automated compared to routine work in process industry, retail and sales. However, several white-collar jobs in finance, sales, marketing and HR are also estimated to be partly or wholly automated, up to 46% of the workforce in these sectors or around 50.000 jobs. No job is really safe, and what was considered unlikely just a few years ago (for example self driving cars) is now a reality.
The Nordics is especially vulnerable to automation since we are industrialised countries with many jobs in manufacturing. However, most jobs likely to be automated are in services, for example online media. When Google launched AdWords in 2001, they established a global online advertising infrastructure with self-serve accounts and no need for sales staff, that in turn created new jobs for online marketers managing ad campaigns. Media agencies now risk being automated as much of the sales and advertising work previously done by humans moves to automated real-time bidding platforms, just as stockbroking was made electronic in the 70s. The owners of the algorithms and platforms that run global online advertising, primarily Google and Facebook, have quickly become the world’s richest companies as the old media giants that still primarily rely on people fight for their survival. The case for automation is just as strong in other industries. So, both manufacturing and services are likely targets for automation. What are the conclusions for the Nordics, or any developed country, to stay ahead in the automation race?
The Swedish Foundation for Strategic Research lists a few areas. First, countries need a strong innovation policy. This includes, for example, resources for research and a positive climate for commercialising new ideas and make innovation flourish. Second, we need policies for lifelong learning, competence development and a modern education system that is designed for the future. Third, we also need to consider tax reforms as employment and traditional income taxes probably cannot be the main base for future tax revenue, as it is today.
The automation of work is not only a way for businesses to become more competitive and profitable. It has, as we have shown, also several effects on the larger society; education, taxation and welfare. It is maybe not surprising that people with deep insights into technology, like Bill Gates, have opinions then about technology’s role in society, and the need, for example, to tax automation to slow down the pace and help people get up to speed with technology.
Elon Musk is another famous tech entrepreneur that is paradoxically worried about the technology development, and its risk for humankind. In Vanity Fair’s feature story “Elon Musk’s Billion-dollar Crusade to Stop the AI Apocalypse” the rebel entrepreneur takes the issue further calling AI and robots humanity’s “biggest existential threat”. He has therefore co-founded OpenAI, a billion-dollar non-profit organisation for safer Artificial Intelligence. One purpose of OpenAI is to distribute AI to the world, rather than let it stay in the hands of powerful organisations and governments.
Elon Musk also has a vision for human-machine cooperation that goes beyond the collaborative robot. Musk’s new company Neuralink wants to turn AI into an extension of the brain. His previous two companies Tesla (electric cars) and SpaceX (space travel) redefine what future humans will do—Neuralink wants to redefine what future humans will be. So if this is a hint about our future, maybe it’s not about robots versus humans, but how how we join forces to work in completely new ways.
In his book Sapiens, A Brief History of Humankind, author Yuval Noah Harari outlined how humans during our 75.000 years of history have shaped the world with innovations like religion, money and empires. It’s a generally positive story about human progress and development. In his new book, Homo Deus, A Brief History of Tomorrow, his outlook is more apocalyptic and he envisions a future where most humans never catch up with the super-intelligence, technology and the resulting job-loss, thereby creating an enormous “useless class”, dominated by a rich techno-elite in power of the algorithms, networks and capital.
This takes us back to the main question. Are robots “stealing” our jobs (a dystopic scenario of the future where we have lost control), or are we humans innovating as we have always done (a positive outlook where we are in charge)? It’s up to us, time will tell. Enjoy your breakfast while it lasts.
Image: Robots deliver dishes to customers at a Robot Restaurant in Harbin, China