Last month, a brief video was uploaded to YouTube that fills me equally with delight and dread. A robot with the physical build of an NFL linebacker and the agility of a wide receiver leaps from a standstill onto a waist-high wooden block. It makes a deft 180-degree hop, then crouches, backflips onto the floor, and raises its arms in triumph. Welcome to the uncanny valley.

Atlas, which its creator, Boston Dynamics, calls “the world’s most dynamic humanoid,” is designed to operate in a human environment. “Atlas’ ability to balance while performing tasks allows it to work in a large volume while occupying only a small footprint,” the company states on its website. “Stereo vision, range sensing and other sensors give Atlas the ability to manipulate objects in its environment and to travel on rough terrain. Atlas keeps its balance when jostled or pushed and can get up if it tips over.”

Atlas is like us, but in important ways, it’s much better than us. I couldn’t make the same leaps. The backflip is totally out of the question.

Machines’ ability to perform human tasks—physical, intellectual, and emotional—improved dramatically this year. “2017 was the year that the robots really, truly arrived,” Wired’s Matt Simon wrote on Tuesday. “They escaped the factory floor and started conquering big cities to deliver Mediterranean food. Self-driving cars swarmed the streets. And even bipedal robots—finally capable of not immediately falling on their faces—strolled out of the lab and into the real world. The machines are here, and it’s an exhilarating time indeed. Like, now Atlas the humanoid robot can do backflips. Backflips.”

But the advances that excite Simon—and surely many other tech bloggers and corporate efficiency experts—have caused widespread anxiety in America in 2017, if the headlines are any indication. “You Will Lose Your Job to a Robot—and Sooner Than You Think,” Mother Jones declared. The New Yorker welcomed “our robot overlords,” adding, “Once, robots assisted human workers. Now it’s the other way around.” The New York Times asked, “Will Robots Take Our Children’s Jobs?” (The worries don’t always pertain to work: “Should Children Form Emotional Bonds With Robots?,” The Atlantic asked.)

Despite decades spent consuming science fiction, we are not at all ready for a society in which we live and work alongside intelligent, autonomous machines. Robot anxiety is real, but it’s more complex than the headlines suggest. The cure is neither to destroy the machines, nor to love them. Ultimately, they aren’t the source of our anxiety. We’re anxious because we’re steeped in a capitalist culture in which our work determines whether we flourish. If we’re all to prosper after the robot revolution, we’ll need to revolutionize our social policies and moral thinking.

Atlas might be our idea of what a robot overlord looks like, but the robots that pose a more direct threat to employment are much more innocuous and decidedly un-human, like the 100,000 of them that run Amazon’s warehouses or those that are taking over fast-food chains. Some even look identical to products that we’re intimately familiar with: Waymo, a subsidiary of Google’s parent company Alphabet, began testing fully autonomous cars—without a human driver as a failsafe—on public roads around Phoenix this year. Waymo’s goal is to launch a driverless ride-sharing service.

And these robots are just the beginning. “By now everyone’s heard the predictions that self-driving cars could lead to 5 million jobs being lost,” Mother Jones’ Kevin Drum wrote, “but few people understand that once artificial-intelligence software is good enough to drive a car, it will be good enough to do a lot of other things too. It won’t be millions of people out of work; it will be tens of millions.” Software that picks stocks is already good enough that BlackRock, the investment firm, laid off dozens of portfolio managers in March and replaced them with algorithms. AI is getting good at medical diagnosis, too. Japanese researchers found that image-recognition software could correctly identify cancerous cells 86 percent of the time. Machines even outperform humans in the crucial visual task of distinguishing between a Chihuahua and a blueberry muffin.

Fear of being replaced can make us forget that robots, like any machine that people build, are meant to make life easier. Even as they’re contributing to layoffs, industrial robots are making some jobs less stressful and more rewarding. The highly automated factories and warehouses that Sheelah Kolhatkar visited for an October New Yorker article are quiet and uncrowded. Some are even dark, because the robots can work without visible light. The relatively few employees of a Michigan office furniture factory say their jobs have become physically much easier as robots took up all of the heavy lifting. The employees will be able to work later into life than they expected when they started at the plant decades ago. But they seem to be the last of a dying breed. As Kolhatkar wrote about a worker at the factory:

He gazed out at the smoothly functioning factory floor, the rows of machines diving and pecking in front of their human minders, performing a kind of dance. Even if the economy stayed strong and demand remained high the head count was expected to decline through attrition, year after year. “It’s got all the technology that you could possibly think of,” Stinson told me, when he showed me the vision table. “Until next week, when we find something else that we could change to make it better.” Automation was bringing greater and greater efficiency, even though, at a certain point, the logic of increasing efficiency would catch up with him, and he wouldn’t be around any longer to witness it. One day, the factory might go dark. In the meantime, he was enjoying the advantages of work that involved less work.

Even as Wall Street firms cut their staffs and factories don’t bother replacing retirees, most American workers in 2017 don’t see an economy where machines are about to take everyone’s job. The unemployment rate is at a 17-year low. We know, vaguely, that machines are getting smarter, but we don’t see quite how they can replicate what we do all day. They don’t participate in the office Secret Santa. So most workers are complacent: A Pew Research Center study found that 77 percent of respondents believed robots would eventually replicate most human work tasks, but only 30 percent thought that robots could do their job.

Faced with Atlas and Amazon bots, our main problem is not that we lack skills. It’s that we lack imagination. The fact that most of us can’t fathom robots replacing us at work suggests that we also can’t imagine the myriad other ways our lives might be different in a more automated world— for better and for worse.

The real source of robot anxiety isn’t the android’s uncanny mimicry of human movement. It’s American capitalism, which forces workers to experience employment as a zero-sum conflict. (When Americans aren’t afraid robots will take our jobs, we’re afraid immigrants will.) We fight hard to keep our jobs because we know that if we’re out of work, we’ll have no way to provide for ourselves or earn social respect. That’s been true for much longer than we have been dreaming of automated production.

Maybe robots won’t replace a third of the American workforce by 2030, as a recent McKinsey study claims. Maybe we will come up with other jobs to do. Or maybe, as some economists have begun to argue, we need more robots because we have too much work to do, and not enough humans to do it. But the past offers sobering precedent. The first industrial revolution created vast wealth and accelerated colonialism. It also reshaped everything from gender roles to the understanding of time in the West. We should expect political and economic changes to result from the new machine age, too.

We will need all the imagination we can muster to craft policies that ensure the robot revolution serves broad human interests. The Pew study revealed that a solid majority supports legislating limits on industrial robots’ use, and some governments are exploring new forms of taxation as companies replace human labor with smart machines. South Korea is instituting what amounts to a tax on industrial robots. A San Francisco city supervisor, Jane Kim, has proposed a similar tax, with revenues going to boost wages or institute a basic income. Drum, of Mother Jones, proposes universal basic income, creative taxation, and socializing ownership of intelligent machines as ways to deal with mass unemployment and inequality in the new machine age.

Any policies we pursue will be progressive ones, simply by default. Drum is right that, as he put it, “The monumental task of dealing with the AI Revolution will be almost entirely up to the political left.” Trump thinks he can revive the manufacturing sector—with human jobs, no less—and the Republican Party is more interested in cutting corporate taxes than guaranteeing a decent living for anyone who isn’t already rich.

Looking ahead to next year, the political left needs to think more seriously about not just the economics of automation and AI, but the social and moral vision of lives lived alongside intelligent, productive machines. It needs to be bold in imagining alternatives to our current capitalist models of education, leisure, and human dignity. The jobs that college students are spending a fortune training for might not exist once they hit midlife. How should universities prepare them for that? Even if we figure out how to distribute machine-made wealth broadly across society, what will free time mean without work to rest from? What’s the value of a human life, if no one needs its labor power?

In short, we need a totally new moral vision for a society that isn’t founded on paid human labor. Pew found that 64 percent of Americans think people will be at a loss for what to do with their lives if robots end up doing most of the work. We’ve put so much value on “productivity” for so long, we’ve forgotten the value of doing other things—or doing nothing at all.

Our narrow, fearful, industrial-age mentality forces us to understand robot anxiety as a problem for individuals: Will a robot replace me? But this is a societal question. The agility and intelligence machines exhibited in 2017 suggests that robots, eventually, will be able to perform your job, no matter how much human creativity you think it requires. And even if that doesn’t happen in your lifetime, all labor will be affected by automation—in ways we can’t even imagine. The challenge will be to transform our individual anxiety into a collective desire for a dignified human future. The robot revolution will flip every one of us upside down. If we start 2018 talking about new models for labor, capital, and the good life, we just might stick the landing.