Humans Rebooted: The Rise of Artificial Intelligence

You’re a 40-year-old accountant with a masters degree, two decades’ experience and a stellar CV. Like everyone you know, you’re also broke, unemployed and living with your parents. The year is 2035 and the robots have landed.

It’s a bleak — and contested — view of the future for professionals. But if the bolder predictions about the march of the machines have any merit, it’s also one to take seriously.

Computers are already capable of a wide range of complex tasks traditionally reserved for highly skilled humans, from diagnosing cancer to monitoring progress on construction sites, and they show no signs of slowing down.

Artificial intelligence (AI) researchers believe robots have a 50% chance of outperforming humans in all tasks in 45 years, and taking all jobs in 120, according to a study by Oxford and Yale Universities published in March. Another widely cited Oxford University paper from 2013 estimated that almost half of jobs in the US were “at risk” of being automated in the next two decades.

Not everyone signs up to the Domesday narrative. An analysis from UK-based consultancy PwC released earlier this year predicted that automation would create new employment and boost productivity, even as it threatens the jobs of 10 million British workers performing routine tasks. While the forecasts vary in outlook, however, they appear to agree on one thing: upheaval in some form is on its way, and some skills are on their way out.

For people who depend on those skills to make a living, this is obviously more than a technical point. And for society more broadly, the rise of the robots raises knotty human challenges. How do we cope with the psychological fallout of mass-scale redundancies? How do we ensure that people can afford to live comfortably if there is less work to go around? And if, in the end, robots do take all our jobs, how do we find meaning in lives without livelihoods?

Coping with obsolescence

Mass, robot-induced obsolescence may not yet have befallen humanity, but lay-offs are hardly a new phenomenon. History is littered with examples of people struggling to cope after finding their services no longer required, offering some clues about the challenges in store — and strategies for managing them.

That includes recent history, explains Dawn Norris, associate professor of sociology at the University of Wisconsin-La Crosse and author of the book Job Loss, Identity and Mental Health. Norris’ study of the psychological battles of US professionals made redundant during the Great Recession has convinced her that society needs to manage advances in AI carefully if we are to avoid widespread mental health problems. 

A key issue, says Norris, is how bound up with our sense of identity paid work is. Take it away and we become vulnerable to various forms of mental distress. Many of the participants in her study reported anxiety, depression and other problems linked to the blow they had taken to their sense of self, she says. “When I asked what the hardest part of job loss was, almost half told me it was their loss of identity. They said, ‘I don’t know who I am any more’.”

If entire professions were to be displaced by robots, the identity crisis could be particularly acute, says Norris, since there would be no obvious remedy. An effective coping mechanism she observed during the recession was for people to find ways of using their professional skills in the short term while they waited for new paid work. A former banking vice-president, for example, found solace in voluntary accounting for his church. But if your skills are simply obsolete, how to cope then?

Part of the answer may come from sharing lessons between cultures, says Ofer Sharone, assistant professor of sociology at the University of Massachusetts. “When you’re looking at automation replacing jobs … I think that societies are differently positioned in their capacity to support people through that transition,”he says.

This view is informed by Sharone’s own comparative research into the experiences of laid-off tech workers in San Francisco and Tel Aviv. This revealed significant differences in US and Israeli responses to unemployment, with implications, he says, for their ability to deal with the fallout.

Like Norris, Sharone found high levels of self-blame among his American subjects, even in the context of wider market shifts, something he ascribes to the “insidious” message sent out by US self-help culture that individuals just need to take control of their lives to succeed. In Israel, where workforce support is less focused on individual responsibility than market analysis, people were more likely to blame the system, he says. While the US professionals were prone to depression and anxiety, the Israelis were more likely to feel angry and betrayed — a form of mental anguish also, but one that allows for greater resilience, suggests Sharone.

“If you think the issue is about the market and it’s external to you it’s not as debilitating,” he says. “You are able to make adjustments. Maybe you’ll try to apply for a new kind of job, or retrain. The difficulty with thinking it’s about you is that that seems unchangeable.”

There are concrete lessons here for how societies can deal with growing numbers of long-term unemployed, says Sharone, who has been working with career coaches in Boston to put these lessons into action at a local level. For one, they’ve ditched the relentless messaging about positive thinking and seizing control in favour of clear facts and figures about the obstacles people face and market conditions. Armed with knowledge, says Sharone, people are better positioned to think strategically about jobs.

Sharone is not alone in advocating an information-based approach to helping humans muddle along with robots. In September, UK innovation agency Nesta released a report on the outlook for skills and employment in 2030 — itself produced in part by machine learning — which provides a granular view of the future for different sectors, and is intended for use by “educators, businesses and governments for strategic and policy-making purposes”.

The report’s sanguine message is that automation will not be a story of obliteration, but adjustment. Some professions, including design and engineering, are strongly complemented by digital technology and will thrive in coming decades, while others, like financial specialisms, will fare less well. Rather than panicking about robots, we should be rationally preparing the workforce for a different skills landscape.

The keep-calm-and-adapt approach is as much grounded in history as future-gazing, explains Harry Armstrong, head of futures at Nesta. Society is stuck in a cycle where every few decades we have “exactly the same conversation” about automation, with dire predictions about mass unemployment that never come to pass, he says. “Partly because it’s such an exciting idea, being able to capture something innately human in something artificial, it captures people’s imaginations and they kind of run with it.”

In fact, workplace evolution is almost always more gradual, he says: “With some occupations you get complete disruption, but that’s pretty rare. What normally happens is you’re getting tasks or aspects of jobs disappearing or shifting into other places … and the occupation then adapts to that, so it takes on other tasks, or it morphs and combines itself with another occupation. It’s only after a period of time that you see there’s been a bigger change.”

The argument that “this time is different” gets short shrift from Armstrong. Apparently people always say that too.

The meaning of work

But what if this time it’s different? There’s at least a case to be made. Robots are already eliminating jobs and depressing wages in the US according to its National Bureau of Economic Research. And their ability to mimic people is improving all the time. A chatbot called Eugene passed the so-called Turing Test back in 2014, successfully tricking at least some members of a panel of judges into thinking it was human. In 2016, a novel co-authored by a Japanese AI managed to get through the first stage of a literary competition.

Tim Dunlop, author of Why the Future is Workless, thinks changes in the workplace will mean more than simply swapping some skills for others. He sees the future of work as “a wicked problem on the same level as climate change”, and believes we are at the least in for a long and challenging transition period.

For Dunlop, the crucial question is not will they or won’t they take our jobs, but how the nature of our daily lives will change when there’s less need for us to work. It’s a question that alarms people, he says, but it’s also one with a potentially positive answer. Passing formal work over to robots could give humans more time to devote to undervalued activities like childcare, he says, with profound implications for how we interact with and judge our fellow humans.

“At the moment there’s paid work and we take that seriously. You have a job and you pay your taxes and that’s the measure of your worth as a citizen,” says Dunlop. “But we completely undervalue the informal economy — looking after kids, aged care, volunteering and community work. Without that, the formal economy of paid work collapses entirely, but we don’t value it.

“As technology integrates itself into the formal economy more, you theoretically open up the possibility of everybody doing less work … So it gives us an opportunity to reassess what we mean by work and what we value in a society.”

The assumption that paid work is fundamental to human experience is a modern construct, Dunlop adds. In the US, the idea of selling your labour was anathema to the founding fathers, who envisaged a nation of yeoman fathers, he says, while in Ancient Greece, “work was so badly thought of that a self-respecting citizen wouldn’t do it”.

The point is not that we should return to the past, but that our attitudes are malleable. “We have to stop thinking in terms of good work and bad work,” he says. “I think once we break down those sorts of barriers, people are more likely to think, actually, I don’t need to prove my worth and gain meaning simply by doing paid work. There’s other ways I can feel fulfilled and contribute to society.”

There’s just one problem. Swapping gruelling hours in the office for time with loved ones may sound lovely, but those gruelling hours are how we pay the bills. If we’re all going to work less, we’ll also need new ways of affording life.

It’s one reason why support is growing for an overhaul of social security, and specifically for the introduction of a Universal Basic Income (UBI), under which all citizens would receive an unconditional stipend each month from the state. A number of countries are experimenting with limited forms of the idea, including Finland. In January 2017, it launched a trial under which 2,000 unemployed citizens receive €560 per month for two years.

One argument is that predictability of income would give people the confidence to seek out new opportunities. Another is that, in a world of higher productivity but lower employment, we need new mechanisms for redistributing national income. Dunlop even suggests such payouts could be considered rightful compensation for the free data we supply to companies like Google and Facebook.

Heikki Hiilamo, professor of social policy at Helsinki University and close observer of the Finnish project, says growing interest in UBI reflects a recognition that existing welfare structures are unfit for the uncertainties of the modern labour market. “This idea of long-term employment with one employer over your whole career doesn’t exist any more. It’s why we need to rethink social protection.” When it comes to middle classes displaced by automation, however, he says it’s too early to tell if it’s the right policy instrument.

There are other concerns too. Anna Coote, principal fellow at the UK-based New Economics Foundation, for example, has argued that a basic income could weaken pressure on employers to provide decent pay and secure jobs, in effect letting them “off the hook”.

Finding our role

The degree to which employers should be on the hook for the pace and scale of change in the workplace is itself a fraught question, and one with warnings from history. During the Industrial Revolution, the state failed to protect the population from the impulses of the new capitalists, says Adrian Randall, emeritus professor of English social history at the University of Birmingham, resulting in friction that might otherwise have been avoided.

“If you go back into the later parts of the 18th century when the major innovations took place … people are far more open to exploring ways in which they can amalgamate new technology rather than simply resisting it. The resistance gets more and more desperate as the employers and innovators use all the devices they can to force through change,” he says. Even the Luddites, who smashed up machines in fear for their jobs, were not the blind opponents of progress we imagine, says Randall. They just wanted change to be incremental.

“I suspect the real lesson we could learn from the past is to have some respect for those whose skills are under threat and to seek ways to integrate them in the new processes of production,” says Randall. “But to do that would involve seeing people not as mere units of production.”

Might technology help us to shift our view? It’s the hope of David Ferguson, head of digital innovation at EDF Energy and member of the advisory board to a UK parliamentary group on AI. He argues that the advance of intelligent machines will require “humans to become more human and to focus on the things we’re actually good at.”

Robots, says Ferguson, are fantastic at doing very specific tasks — EDF now uses an AI to spot seaweed or jellyfish blooms that may get sucked into the cooling water intake of coastal power stations, for example. But when it comes to complex situations requiring empathy, understanding and lateral thinking, he says, humans are still your man.

In the longer term that has implications for company structure and culture, he suggests. In customer services, for instance, where EDF currently has 3,000 people manning the phones, it’s likely there’ll eventually be a smaller number of people performing more valuable tasks — solving problems, caring for customers and sorting out difficult situations rather than asking people to read out their account number and pointing them to an answer on a website. That in turn could require a less regimented management approach, says Ferguson. “You’re going to have to put more trust in your people and let them take responsibility for outcomes rather than just following specific steps.”

Even if companies reorganise themselves for an AI world, however, they still face an overarching challenge: if there is mass unemployment, who will buy their products or services? For Ferguson, the societal question is urgent, but not one companies can tackle alone. “Government absolutely needs to lead the way,” he says. “And we need this to happen really quickly because the speed of progress in AI is exceptionally fast. “Things are going to change quite quickly.”