By John Gapper, Work.
On 9 March 2016, an event occurred with profound implications for game-playing, leisure and the future of employment in the 21st century. Lee Sedol, the South Korean world champion of Go, an ancient Chinese two-player strategy board game, resigned after 186 moves in his first match against AlphaGo, a computer program created by DeepMind, the UK-based artificial intelligence (AI) start-up owned by Google.
Lee, who went on to lose to AlphaGo by four games to one, had predicted that the big question would be the scale of his victory — would it be 5-0 or 4-1? Yet AlphaGo had been learning so rapidly, playing millions of games of Go against itself, it had achieved a feat that was expected to take a machine several decades longer.
Computers and robots are already doing surprising and impressive things. A multi-tasking bot from Momentum Machines can make — and flip — a gourmet hamburger in 10 seconds. Associated Press has already hired robots to write financial reports. And Google has filed a patent to start making robot workers with distinct personalities.
All this raises serious questions about the future employment of millions of people. If machines can do all this now, what talents could they master in 30 years’ time? What will be left to human employees when machines are so physically and intellectually dexterous?
“AI is attacking the heart of what humans have always believed is unique to us,” says Toby Peyton-Jones, Siemens’ UK human resources director. “We never believed that machines would be able to drive cars or do a better job than doctors, but they have learned and they keep on learning. They can master in a matter of hours things that take us 20 years to learn.”
Challenge to One Education for One Career Model
Technology is not the only disruptive force reshaping the future of employment: human longevity is another. As we approach what Lynda Gratton and Andrew Scott, two professors at London Business School, have dubbed “the 100-year life”, assumptions about the typical career are being upended. By 2064, according to the Office for National Statistics, the life expectancy for English women is projected to reach a century. Leaving education in your twenties and retiring at 65, often after sticking to one profession, even one employer, for an entire career, may no longer be possible.
While people’s lives get longer, company lifetimes are shortening. People will not work for the same company throughout their career partly because the business itself may not endure as long as its staff. Corporate turn-over is rising amid globalisation and intense competition; one study found that the average firm in the US Standard & Poor’s 500 index in 2012 stayed for only 18 years before dropping out. As Tracey Wilen, a former technology executive and visiting scholar at Stanford University, says: “Most companies do not last more than 15 years. We will outlast the company, so it cannot take care of us.”
Such shifts will require workers to adapt, taking more responsibility for their careers and skills, but governments and employers will be under pressure to reform education and training. We assume young people can gain a thorough education by their early twenties, which will largely see them through their entire careers. They may top that up with training, but they don’t need to start again in mid-life. The combination of advancing technology and longevity brings that model into question.
Reform is not just a matter of extending training over a lifetime. The very nature of education is in question. At the moment, a teacher delivers a lesson to individual students on standard subjects. Yet the qualifications currently most in demand — science, technology and mathematics — contain the kind of structured information that computers are best at learning.
Evolution of Demand
“You can anticipate the evolution of labour demand, even if the predictions are not exact,” says Andreas Schleicher, Director for Education and Skills at the Organisation for Economic Co-operation and Development (OECD). “Digitisation affects the routine cognitive skills that computers are best at. There will be a decline in things that are easy to teach and easy to test. Unfortunately, that is where education focuses.”
Concerns about skills becoming technologically obsolete go back to the early 19th century when the Luddites, a group of English textile artisans, destroyed looms to protest the loss of their jobs. In 1964, US President Lyndon Johnson appointed a panel to study “technology, automation and economic progress” amid fears that rising productivity would cause unemployment.
The panel concluded that the general level of demand for goods and services had more influence on employment levels than technology. But it took the threat seriously enough to recommend an idea that has returned to favour: a universal basic income for families who could not find work. It also wanted two years of free post-school education at community or vocational colleges, and the government to become an “employer of last resort”.
Half a century later, unemployment is still low in both the US and UK, but some worries about the erosion of skilled work have come true. The US underwent two striking changes in employment in the 20th century. First, it shifted from agriculture to industry — a trend mirrored in China after Deng Xiaoping’s economic liberalisation in the 1980s. Some 41% of the US workforce was employed in agriculture in 1900, but only 2% a century later, so there was much less physically demanding, repetitive, risky and boring work around.
The second major shift was less propitious. In the second half of the 20th century, US employment “hollowed out” as jobs for mid-level manual and clerical workers — many of them male — were lost to automation, and manufacturing shifted to China and low-wage countries. Skilled blue-collar and white-collar jobs were lost, and in their place came lower-wage personal services and “McJobs” at fast-food restaurants and courier services. The economic returns for university and professional careers became higher.
This trend holds — education is still vital to obtain a good job, and university and postgraduate education, especially in professions such as finance and law, remain hugely valuable. Yet there are big questions about whether, or how long, that will last.
Exponential Advances in AI
The advance of technology, particularly machine-learning and AI, could take over jobs — and skills — previously regarded as safe. Instead of peering at job destruction from a safe height, professionals will experience it themselves. In their book, The Future of the Professions, Richard and Daniel Susskind argue that society “will neither need nor want” professionals to work as they now do, and that “increasingly capable machines” will manage much of their work.
Two aspects of modern computing mark it out from the technology employers and workers have confronted in the past. One is an ability to interact with the human world in a more sophisticated manner than before. AI is making advances in facial and voice recognition: computers can already interact with people in a basic manner and will eventually be able to work as part of teams.
The second crucial aspect of modern computing is AI, or machine-learning. AlphaGo’s defeat of Lee was surprising because Go is a complex game. Players have to “sense” the board — to have a positional understanding that goes beyond a clear set of principles. This is known as “Polanyi’s paradox” after the Hungarian-British polymath Michael Polanyi, who wrote that much human knowledge is tacit, and that “we know more than we can tell”. That makes it difficult for any algorithm to imitate.
In effect, AlphaGo has learned by playing in a human-like fashion. This promises to take computers beyond their existing limits in professional work, such as analysing millions of legal documents by following a set of rules laid down by professional lawyers. They may be able to adapt and learn more flexibly, performing a wider variety of accounting and consulting work.
In the future, some of us could be working as…
As cities get greener, someone will have to tend to the new infrastructure. Enter the urban shepherd, who will take care of a city’s beehives, ensure that people manage compost correctly and help to curate the growing number of vertical gardens.
Michael Crichton’s Jurassic Park vision could almost come true. We could see a niche profession for geneticists who reintroduce extinct species. It’s too late for dinosaur DNA to be of use, but there’s hope for mammoths or dodos.
Ever feel that the technology that is designed to make you work smarter is wasting your time? Fear not. You’ll be able to pay a productivity guru to help you exploit the technology. They may even refer you to a digital detox therapist.
This concludes Part 1 of the article “Your Services Are No Longer Required”, first published in the Summer 2016 issue of CIPD’s Work. magazine. Part 2 of this article can be viewed here.