A World Without Work summary

You’ve certainly heard all kinds of outlandish predictions about how technology will change the world. Our societies will never be the same again! We will all become redundant! An army of robots will take our jobs! 

Optimists and pessimists agree that there needs to be a change of direction. But what should it look like?

These blinks will cut through the noise and give you a better understanding of what exactly automation is, what it’ll do to human societies, and, most importantly, how we can embrace it to build a better world. 

You’ll draw on lessons from history and economics to learn the history of technological change, and you’ll also see a new path forward – a path toward a world where work isn’t necessary for everyone to lead happy and fulfilling lives.

In these blinks, you’ll learn

• why we can’t know what the future holds;
• some ways to bridge the great divide; and
• that Homer wrote about AI – in a manner of speaking.

Whose jobs are machines taking? Is it people who work on assembly lines? Or cashiers at supermarkets? Or should brain surgeons, too, worry that one day a robot might replace them?

Well, the growth of technology will certainly affect everyone. But recent trends suggest some clues as to which sectors of the economy are more likely to be automated.

The key message in this blink is: All jobs are at risk from technological change.

Over the last few decades, technology benefited highly skilled, formally educated workers more than their low-skilled neighbors. Why? Well, the answer is: computers. Between 1950 and 2000, their power increased by a factor of ten billion. 

This created a demand for highly skilled workers who could operate new machines. As demand grew, so did supply – and more and more people learned to use computers. Predictably, that pushed wages down. But then something interesting happened: demand continued to grow, and the wages of highly skilled workers actually began climbing. By 2008, economists were recording an unprecedented income gap between US college graduates and people who’d only completed high school.

So, does this mean that technology always offers more benefits to well-educated people? Not really. In the past, it was actually the other way round. Remember the luddites? In eighteenth-century England, weaving was a high-skill job. When mechanical looms came onto the scene, you no longer needed advanced training to make good cloth. Low-skilled workers reaped the rewards.

So who’s going to benefit from automation in the future? Economists believe that technology is boosting both low- and high-skilled jobs; it’s the middle class that suffers. There are more cleaners and lawyers, but fewer secretaries and salespeople. 

One explanation for this was put forward by a trio of MIT economists. Their theory says “routine” tasks are easier to automate than “non-routine” work – the sort that depends on creativity, judgment, interpersonal skills, or complex manual work. 

Simple, routine skills can be explained, broken down, and turned into algorithms. Computers have little trouble with them. But machines struggle with non-routine skills, which are more complex and harder to explain. 

For a long time, people believed that non-routine jobs were safe from automation because scientists couldn’t teach computers to do them. However, that’s beginning to change. As we’ll see in the next blink, machines have already begun to teach themselves.

The ancient Greek poet Homer is most famous for The Iliad and The Odyssey. But did you know that he wrote about more than just heroes and battles? He also described what we would now call artificial intelligence, or AI. In The Iliad, Homer tells of “driverless” three-legged stools that would come to their owner when called – a bit like today’s autonomous cars. 

Now, maybe Homer didn’t have self-driving vehicles in mind when he told that story. But it speaks to an important point: humans have long dreamed about machines capable of autonomy. And recent breakthroughs are making those dreams a reality. 

To understand what AI can do, let’s begin by looking at its history. The first attempts to create artificial intelligence date back to the mid-twentieth century, when computing was coming into its own. Early AI researchers attempted to replicate human thinking. 

To develop chess software, for example, they asked grandmasters to explain their thoughts about the game. And then engineers tried to teach the computer those same processes. 

The key message of this blink is: The breakthrough in AI research came when computers stopped trying to think like humans.

By the late 1980s, however, this approach had stalled. Whether it was chess, translation, or identifying objects, early AI couldn’t beat a human by thinking like a human. So what was the solution? Scientists realized they needed to change tack – imitating human thought could only get computers so far.

The next wave of AI research was based on a more pragmatic approach. Scientists gave machines a task and told software to accomplish it in whatever way possible – even if it didn’t make sense to human beings. Instead of using traditional chess or translation strategies, new AI programs took in hundreds of millions of data points and scanned them to find patterns. 

As a result, the field of AI research took a giant leap forward. In 1997, IBM’s Deep Blue beat world chess champion Garry Kasparov. And AI isn’t just excelling at chess; modern image identification programs, for example, routinely outperform humans in competitions.

These advances are critical to understanding how automation might affect the future of work. Economists used to think that computers would never function without human guidance. But machines can now find nonhuman solutions to problems and tasks. And this suggests that machines will someday be able to learn non-routine skills, which were previously considered beyond their capabilities.

Science-fiction author William Gibson once said, “The future is here – it’s just not evenly distributed.” It’s useful to keep this line in mind when we’re discussing automation. When it comes to what computers can do, the future is very much now. AI can outperform humans in an ever-increasing number of tasks, from detecting liars to manufacturing prosthetic limbs.

But the jump from “AI can do it” to “AI will do it” is a big one, and this gap differs from one country to the next. 

The key message here is: Machines are getting better at all kinds of jobs, but technological progress will look different everywhere.

As technological capabilities continue to increase, automation will transform every single industry. Just look at agriculture! Today’s farmers have driverless tractors, facial-recognition systems for cattle, and automatic sprayers. In Japan, 90 percent of crop-spraying is done by drones. Even tasks that previously took fine motor skills have been automated – robots can pick oranges by shaking them off trees.

What about industries that require more complex thinking? Well, law, finance, and medicine have all seen the arrival of software that can analyze more information than us humans. AI is great at seeking out relevant patterns and past cases. 

A diagnostic system developed by Chinese tech firm Tencent in collaboration with a Guangzhou hospital draws from more than 300 million medical records to evaluate patients. 

Even jobs that require feelings and emotions can now be done by machines. Some facial recognition systems already outperform humans when it comes to telling whether a smile is real or not. 

“Social robots,” so named because they can detect and react to human emotions, are expected to become a $67 billion industry. They’re increasingly common in health care; a humanoid machine called “Pepper” has become a resident of some Belgian hospitals. Its job is to greet patients and escort them around the warren of corridors and buildings.

But just because more tasks can be automated doesn’t mean they will be. Automation requirements and costs vary across regions, so countries won’t necessarily move at the same pace. 

For example, Japan has a lot of elderly people but lacks nurses. So hospitals there have a strong incentive to automate care work. In contrast, countries with younger populations – and lots of people willing to take on low-wage jobs – have less incentive to automate health care. Indeed, there may even be political pressure to stave off the arrival of robo-medics.

Looking for a job isn’t much fun, is it? Now imagine that you’re unemployed because your job has been automated. How are you supposed to find a job if you’ve been replaced by a machine? This challenge will most likely affect millions of people. 

In the previous blinks, we learned that automation will expand the size of the economic pie. In other words, machines will create new jobs, which will offset unemployment. But will people fill those new jobs? Well, the answer is far from straightforward. There are too many obstacles along the way – like, for example, the mismatch of skills. 

The key message in this blink is: Increasingly capable machines will lead to huge job losses. 

Suppose most new jobs are high-skilled – things like specialists in AI management. That would be of no help to low-skilled factory workers.

Or consider the geographic mismatch. Would you want to move hundreds of miles for the sake of a job? The internet has helped with this by making remote work more feasible, but geography still matters. Think of Silicon Valley, which attracts so many tech companies. The number of talented programmers in the region means that start-up founders often move to Silicon Valley to find talent and form new connections.

Economists call such obstacles “frictions.” Scientists believe that they’re all short-term – that they’ll smooth out in the long run. But there’s something that won’t go away: structural change of the labor market. 

We’ve seen that technology boosts productivity and overall output. As technology develops, though, it will reach a point where it won’t need humans at all. Think of the cars, for example. First, GPS complemented humans by giving taxi drivers the ability to take more efficient routes. Nowadays, however, driverless cars are set to replace them entirely. Even as demand for taxis increases, it won’t mean more jobs for humans; companies will just produce more driverless cars. 

This shift won’t happen overnight. As Silicon Valley researcher Roy Amara once said, “We tend to overestimate the effect of a technology in the short run, and underestimate the effect in the long run.”

But how long is long? Well, these effects will be seen in decades – not centuries – and they’ll continue to accelerate as AI gets smarter. The trend is clear: even though output will continue to increase, there’ll just be less work for humans to do.

For most of human history, we struggled for subsistence. This term, coined by the well-known economist John Maynard Keynes, meant that human societies simply weren’t producing enough for everyone to live on. This made the question of how to distribute resources an afterthought.

Today’s technological advances mean that we manufacture enough for the entire world to live comfortably. The economic pie has grown in size – but how should we divide it? 

If we look at the economic data, one thing becomes glaringly clear: in recent years, slices of the pie have become less equal. 

The key message of this blink is: Automation has increased inequality by widening the income gap between jobs.

Let’s think about the stuff we own – our capital, in economic terms. We can divide our capital into two categories: traditional capital and human capital. Traditional capital is stuff you can buy like land, equipment, or intellectual property. Human capital is a broader concept that encompasses all of your skills and abilities. 

If everyone had plenty of traditional capital, automation wouldn’t be a big deal. However, most people control very little traditional capital. Instead, they use their human capital to generate wealth. When their job becomes automated out of existence, they lose that human capital. See the problem?

Data shows exactly how big of an issue this has become. Before 1980, income growth was steady for all Americans. Between 1980 and 2014, however, things changed. Low earners enjoyed little, if any, income growth. But the 1 percent of the population that were already earning the most saw their income skyrocket.

This pattern has been repeated in wealthy countries all over the world. For all but the very highest-skilled workers, human capital just isn’t worth the same as before. The end result? In the USA, the poorest 50 percent of Americans own only 2 percent of the country’s wealth, while the richest 1 percent own 40 percent.

There are a couple of important takeaways here. The obvious one is that changes in the future of work will lead to soaring inequality. Another question is: How will society function when people no longer need to work?

Traditionally, most of us have had to work for our share of the economic pie. But what happens when automation eliminates jobs? How are societies supposed to support people who have been replaced by machines?

If the labor market won’t do it, another institution has to step up. There’s only one organization with the power to do so: the state.

The key message here is: When automation collapses the labor market, the “Big State” must ensure that wealth is distributed.

Most developed countries already have a welfare state. This concept, in its modern form, dates back to the beginning of the twentieth century. 

But the welfare state, too, needs to change. It was originally designed to supplement the labor market. The fundamental principle was that people who work support those who don’t. Employment was the expectation.

But in the automated world, this will no longer be the case. The welfare state will have to be replaced by what the author calls the “Big State” – an institution that understands there won’t be enough work for everyone.

The Big State will have to accomplish two primary goals: tax those who benefit from automation, and redistribute income to those who are harmed by it. 

The Big State can tax workers like software developers or tech company managers. And it can also collect money from owners of traditional capital including land, machines, or property rights. Finally, this state can tax businesses – especially those receiving extra profits because of all the automation they use. 

Once the Big State has gathered all this cash, how should it distribute the money? Some economists have floated the idea of a Universal Basic Income, or UBI, which would provide cash payments to everyone.

But the author thinks this idea can be tweaked even further. He proposes a Conditional Basic Income, or CBI. This payment would support a specific community. 

CBI avoids one of the drawbacks of Universal Basic Income – its perceived unfairness. If everybody got money from the state, this would feel unfair to some people; we’d face the risk of split communities and even conflict. 

In contrast, the author’s proposed CBI would only go to people who met certain criteria. This system would enable earners to share their wealth with individuals they really wanted to help. 

In this way, a better and more stable society could emerge – one where people had less work, yes, but still felt support from strong communities around them.