by Andrew Yang
Most people assume that humans will always have the advantage over AI when it comes to work that requires creativity, like painting or music, and jobs that requires nuanced, sensitive human interaction, like therapy. In fact, Google’s neural network, a computer system modeled to “think” like a human, has produced art that could easily be confused for a human being’s, like the work on the next page. You can also check out a symphony online that was composed by a software program, Iamus, which many listeners found indiscernible from a human composition when it was performed. Google “Adsum” by Iamus and take a listen.
You might have figured that therapy would be the last province of automation. If so, you were wrong. USC researchers funded by the Department of Defense in 2016 created an AI therapist named Ellie to treat veterans for post-traumatic stress disorder (PTSD). Ellie appears as a video avatar and provides soothing questions and responses. Ellie measures voice tone and facial expressions to try to identify whether a soldier needs to seek additional treatment with a human counselor. Early research is promising and indicates that soldiers often feel more comfortable confiding in a clearly artificial therapist than an actual human being. Ellie is meant to be a complement to human therapists—but one can easily imagine her checking in with patients in between appointments and taking on more over time.
When I was 13, I had to have four teeth pulled in preparation for wearing braces. I was actually kind of excited about it because I saw my dad’s teeth and was like, “whatever it takes, let’s not have those.” I remember going to the dentist and wondering what kind of magic he would employ to pull the teeth. Not much—Dr. Goodman just put some pliers on the first tooth and yanked and jerked until it came out. The second one was stubborn and he had to shift positions a few times—I remember him putting his foot on my chest and yanking away.
I walked away thinking, “Wow, dentists have to be kind of strong to do what they do.” Also, my jaw hurt.
I tell this story because often the boundary between what we consider intellectual and manual work will be unclear. Surgeons are among the highest-trained, most highly compensated doctors because cutting people open is a big deal. Yet their highest-value work is, for the most part, manual and mechanical. My surgeon friends often swear off activities like basketball because they are worried they’ll hurt their fingers or hands.
Some jobs might not go away the instant new technology arrives that could replace them. Much of how automation unfolds in medicine is dependent upon regulations and licensing. It is presently illegal to do many things without a doctor or pharmacists’ license. This is very likely to be a field where technical innovation far outstrips implementation because doctors will fight the steps, and they have a very powerful lobby. They will argue that no one is as good for a patient as a highly trained human doctor, even in the face of dramatic evidence to the contrary as AI improves. Some patients also might prefer seeing a human doctor, though I suspect this preference will fade over time.
There are many obstacles to AI truly becoming broadly intelligent—one neuroscientist described most systems today as being better than a human could ever be at one specific task and dumber than a two-year-old at anything else. Still, our conception of what is beyond the capacity of a computer is about to change. There is a lot of white-collar and creative work that can be automated. In startups, we have a saying of what to do when you’re not sure what the answer is: “Throw money at the problem.” Soon, the answer to everything will be “Throw AI at the problem.”
If you think your job is safe from computers, you’ll probably be wrong eventually. The purpose and nature of work is going to change a lot in the next 10 years. The question is what will drive this change aside from the fact that fewer of us will have jobs to go to.
SEVEN
ON HUMANITY AND WORK
I’ve been in one significant car accident. I was 20 years old, driving at night from Providence to visit my brother in Boston. I was behind the wheel of my family’s old Honda Accord. It was a rainy night. As I approached Boston, full-speed highway traffic ended with a traffic light, and I noticed only when the stopped car in front of me was way too close. I jammed the brakes and my tires screeched, but I still hit the car in front of me hard. The impact crumpled the rear third of the car in front of me and caved in the front of my old Accord, which folded up like an accordion. I jerked forward into my seat belt, stunned.
After a few seconds I got out and went to the car in front of me. “Is everyone okay?” The sight of the ruined car in front of me made me cringe. No one was hurt—there were three people in the car not much older than me. They were shaken up but fine. They weren’t angry at all. I apologized several times. I felt like a grade-A jerk.
We all waited for a police car and tow trucks as cars went past us in the rain. We made small talk about how we weren’t sure if our cars were salvageable. It took about 30 minutes but it felt like hours. I rode in the passenger seat of the tow truck to the garage and waited there for my brother to come and get me. The garage was closed so after the truck driver left I waited outside in the rain on the curb with my head in my hands.
I remember this night in part because I had broken up with my college girlfriend—or she had broken up with me—earlier that day. This was back when people dated in college. I was upset about it and was heading to Boston to hole up with my brother. It’s safe to say that my emotional state contributed to my inattentiveness, and may even have been a key factor in my rear-ending that car.
Our humanity is what makes us unique. People are the most important aspect of all of our lives.
That said, our human qualities may not always make us ideal drivers, counselors, servers, salespeople, helpdesk workers, and so on. Drivers lose concentration. Counselors break confidences. Servers have bad days and are rude. Salespeople have biases and act inappropriately. Helpdesk workers get bored. And so on. There’s a big distinction between humans as humans and humans as workers. The former are indispensable. The latter may not be.
Yuval Harari in Homo Deus makes the point that our cab driver can look into the sky, contemplate the meaning of life, tear up at the sounds of an opera, and generally do a million things that a robot driver cannot. But most of those things don’t matter to us when we get into the back of the cab. Oftentimes, we’d prefer to be left alone rather than make conversation. I know I’m occasionally guilty of this.
One of the common themes of the new economy is that women are better equipped to excel in the growth areas and opportunities in a service economy, including nurturing and teaching other people, which are among the toughest activities to automate. Conventionally male-dominated jobs like manufacturing, warehouse shelving, and truck driving are among the easiest. I’ve heard women say, “Why don’t men just adapt and take on more ‘feminine’ roles?” That’s a lot easier said than done, and I’m not convinced asking people to go against type because the market demands it is the right response. The market doesn’t care what’s best for us—trying to reshape humanity to meet its demands may not be the answer. On other fronts, there are significant initiatives to include more women in high-paying fields like technology and finance that remain predominantly male.
I’ve started a few companies and enjoy nothing more than building great teams of people who are happy and engaged with their work. That said, I think that many people both overestimate the qualities that humans bring to work and underestimate the drawbacks of hiring humans. Here’s a partial list of things that can make people imperfect workers and management an all-consuming role:
• People generally require a degree of training.
• We typically want more over time.
• We need to rest.
• We require health care that you sometimes must pay for and we can be very particular about.
• We get sick.
• We want to feel good about what we’re doing.
• We have bad days.
• We can’t do the same task precisely the same way mill
ions of times.
• We have families who we want to spend time with.
• We are sometimes bad at our jobs and need to be fired. In which case we generally want severance pay or we will make you feel bad.
• We get bored.
• We have legal protections. We occasionally sue our employers.
• We can become demoralized and unproductive.
• We take 15–20 years of rearing to become productive and then we are unproductive and infirm for 10–15 years at the back ends of our lives. We often want you to pay us to account for both the time at the end and the cost of raising our children.
• If something bad happens to one of us, the others notice.
• We occasionally harass each other or sleep with each other.
• We sleep.
• We sometimes are dishonest and even steal.
• We occasionally quit and look for other jobs.
• We see things. We share information.
• Some of us use drugs.
• We get injured and disabled.
• We are unreliable. We sometimes change our minds.
• We sometimes take breaks when we should be working.
• We sometimes organize and negotiate for various benefits beyond what we could get on our own.
• We sometimes have bad judgment and can act in ways that will tarnish your brand.
• We have social media accounts.
• We expect time off for holidays.
• We sometimes get divorced or have relationships end which can make us sad and unproductive.
• We sometimes talk to reporters.
• You cannot sell us to another firm.
• We do not come with warranties.
• Our software often does not upgrade easily.
At the beginning of the book, I quote Terry Gou, the founder of the Taiwanese manufacturing company Foxconn, comparing humans to animals. He brought 300,000 robots into his factories to supplement his million workers making Apple products in part because 14 Foxconn workers had committed suicide in the previous two years. In contrast, robots do not experience emotions and do not get depressed.
The automation wave is coming in part because, if your sole goal is to get work done, people are much trickier to deal with than machines. Acknowledging this is not a bad thing—it is a necessary step toward finding solutions. It may push us to think more deeply about what makes humanity valuable.
It’s worth considering whether humans are not actually best suited for many forms of work. Consider also the reverse: Are most forms of work ideal for humans? That is, if we’re not good for work, is work good for us?
Voltaire wrote that “Work keeps at bay three great evils: boredom, vice, and need.” The total absence of work is demonstrably a bad thing for most people. Long-term unemployment is presently one of the most destructive things that can happen to a person—happiness levels tank and never recover. One 2010 study by a group of German researchers suggests that it’s worse over time for life satisfaction than the death of a spouse or permanent injury. “There is a loss of status, a general malaise and demoralization, which appears somatically or psychologically” with prolonged unemployment, said Ralph Catalano, a public health professor at UC Berkeley.
On the other hand, most people don’t actually like their jobs. According to Gallup, only 13 percent of workers worldwide report being engaged with their jobs. The numbers are a little better in America, with 32 percent saying they were engaged with their work in 2015. Still, that means that more than two-thirds of Americans aren’t exactly skipping on their way to and from the office each day.
As comedian Drew Carey put it, “Oh, you hate your job? Why didn’t you say so? There’s a support group for that. It’s called everybody, and they meet at the bar.” Most of us struggle to find work that we’re excited about, particularly if we have financial goals and pressures to meet. Even the successful among us have made a number of compromises and learned to more effectively adapt to them over time. When you encounter someone who really likes his or her job, you remember it because it’s so rare.
The relationship between humanity and work involves money, but in something of a negative correlation. The jobs and roles that are the most human and would naturally be most attractive tend to pay nothing or close to nothing. Mother, father, artist, writer, musician, coach, teacher, storyteller, nurturer, counselor, dancer, poet, philosopher, journalist—these roles often are either unpaid or pay so little that it is difficult to survive or thrive in many environments. Many of these roles have high positive social impacts that are ignored by the market.
On the other hand, the most lucrative jobs tend to be the most inorganic. Corporate lawyers, technologists, financiers, traders, management consultants, and the like assume a high degree of efficiency. The more that a person can submerge one’s humanity to the logic of the marketplace, the higher the reward. Part of this understanding in America is a high level of commitment to work—educated Americans are working longer hours than they did 30 years ago, and many are expected to be available via email on nights and weekends, even as working hours have dropped in other developed countries. Four in 10 Americans reported working more than 50 hours per week in a recent Gallup survey.
This wasn’t always the case; Americans workweeks were actually getting shorter up until 1980. John Maynard Keynes, the British economist, famously predicted in 1930 that, given the continued growth in productivity and progress, by 2030 the Western standard of living would be four times higher and we would be working only 15 hours per week. He was right on the standard of living and very wrong on our work hours. Meanwhile, numerous studies have shown that a lot of the work we’re doing isn’t really adding value, and that we could cut our hours and maintain most of our productivity.
Benjamin Hunnicutt, a historian at the University of Iowa, argues that if a cashier’s job were a video game, we would call it completely mindless and the worst game ever designed. But if it’s called a job, politicians praise it as dignified and meaningful. Hunnicutt observes that “Purpose, meaning, identity, fulfillment, creativity, autonomy—all these things that positive psychology has shown us to be necessary for well-being are absent in the average job.” Most jobs today are a means for survival. Without their structure and support, people suffer psychologically and socially, as well as financially and even physically.
Whether work is good for humans depends a bit on your point of view. We don’t like it and we’re almost certainly getting too much of it. But we don’t know what to do with ourselves without it. Oscar Wilde wrote, “Work is the refuge of people who have nothing better to do.” Unfortunately that may describe the vast majority of us.
The challenge we must overcome is that humans need work more than work needs us.
EIGHT
THE USUAL OBJECTIONS
I have had hundreds of conversations about the impact of automation on the labor market with people all over the country of various backgrounds. The most common question I hear is, “If this were happening, wouldn’t we know about it?” People are uncertain and skeptical as to what’s happening with the American economy. A lot of people want to believe what they see right in front of them or what they hear from partisan websites and social media that reinforce their current ideas. It’s hard to believe in things that are developed hundreds of miles away on the campuses of technology companies, often behind closed doors.
Before we move on to the next part of the book and explore what the future might hold, I want to tackle a few of the common questions I hear.
“Aren’t fears of disappearing jobs something that people claim periodically, like with both the agricultural and industrial revolution, and it’s always wrong?”
It’s true that agriculture went from 40 percent of the workforce in 1900 to 2 percent in 2017 and we nonetheless managed to both grow more food and create many wondrous new jobs during that time. It’s also true that service-sector jobs multiplied in many unfor
eseen ways and absorbed most of the workforce after the Industrial Revolution. People sounded the alarm of automation destroying jobs in the 19th century—the Luddites destroying textile mills in England being the most famous—as well as in the 1920s and the 1960s, and they’ve always been wildly off the mark. Betting against new jobs has been completely ill-founded at every point in the past.
So why is this time different?
Essentially, the technology in question is more diverse and being implemented more broadly over a larger number of economic sectors at a faster pace than during any previous time. The advent of big farms, tractors, factories, assembly lines, and personal computers, while each a very big deal for the labor market, were orders of magnitude less revolutionary than advancements like artificial intelligence, machine learning, self-driving vehicles, advanced robotics, smartphones, drones, 3D printing, virtual and augmented reality, the Internet of things, genomics, digital currencies, and nanotechnology. These changes affect a multitude of industries that each employ millions of people. The speed, breadth, impact, and nature of the changes are considerably more dramatic than anything that has come before.
It is true that this would be the first time that the labor market did not meaningfully adapt and adjust. But Ben Bernanke, the former head of the Federal Reserve, said in May 2017, “You have to recognize realistically that AI is qualitatively different from an internal combustion engine in that it was always the case that human imagination, creativity, social interaction, those things were unique to humans and couldn’t be replicated by machines. We are coming close to the point where not only cashiers but surgeons might be at least partially replaced by AI.” Fifty-eight percent of cross-sector experts polled by Bloomberg in 2017 agreed with the statement “it is in fact different this time” and that the labor market disruptions will be severe and unprecedented. The consensus is growing.