Job automation will sweep through the economy at an unprecedented pace in the coming decade.
Will this surge in automation cast tens of millions of fired workers into a newly impoverished underclass? Can enough new jobs be created so we at least muddle through?
Will strong insurgent unions ensure that the economic benefits of increased productivity are broadly shared? Or, to push the envelope even farther, what if the means of production were put under public control and the economy was organized around meeting human needs?
The extent to which we as workers organize ourselves to defend our rights will go a long way in shaping that future.
A 2019 Brookings Institute report projected that 52 million U.S. jobs would be affected by algorithms by 2030. In 2021, global consulting firm McKinsey & Company predicted that 45 million jobs will be lost to algorithms and androids by 2030.
A 2022 analysis by Finance Online reports that 43% of employers are set on cutting down their workforces to make way for advanced technology. The group says that U.S. workers between the ages of 18 to 34 are the most affected by job displacement from automation, that by the late 2020s, nearly 25% of women workers are at risk of being displaced compared to more than 15% of men and that highly-educated people have more opportunities to work in sectors where they are less likely to be displaced due to technology integration.
We’ve already seen automation firsthand. Think self-checkout at the grocery store or McDonalds, or being forced to interact with automated customer-service phone lines and chatbots. Workers at Amazon, Tesla and other Big Tech companies have experienced “cobots” — large robots that work “with” them to get the job done. This often leads to additional work stress, from the knowledge that more technology at the workplace means worker surveillance — the “digital whip” — to the fact that robots are often not programmed to properly understand the pace of the human they’re working with and break frequently, leading to unsafe working conditions in the name of “productivity.”
Consider this testimony from Thometra Robinson, a worker at an Amazon warehouse in Stone Mountain, Georgia:
You want us to perform our job with ineffective, inadequate equipment. People are hurt. I got pictures of piss in bottles and cups. … You gotta tape up your monitor that literally just hangs there, is barely lit up. The scanner, God, Lord, you’re playing with that all day. The belt is always breaking. Mind you, you have to pack a box in 37 seconds, which is just unrealistic. All the training they tell you to do, they don’t tell the computer that. So the computer doesn’t know you gotta do all these little steps. If you’re missing an item, you need to go look for it. You gotta get all the way down on your knees. And there are techniques and protocols that they add. … You add on this extra shit but you still gotta do it in 37 seconds. … You can’t be safe and productive. It ain’t gonna happen. … You have to stock your station; there’s not enough water spiders. But the computer doesn’t know you have to get your own stuff. They just know that you have to pack your box in 37 seconds.
Automation is a recurring feature of capitalism as bosses seek to exponentially increase productivity while reducing the percentage of their revenues they pay out as wages. At the same time, new industries and new jobs are created and unemployment rates remain relatively stable over time, though the process of change can be wrenching for some communities.
With the development of Artificial Intelligence (AI) — computer systems that are able to perform tasks that normally require human intelligence — technology-induced mass unemployment is more likely than in the past. AI technology is being developed to drive cars, clean schools, perform surgery, write articles, compose songs, create digital art and beyond. Some (very rich) people are using AI to digitally preserve their consciousness so that replicas of themselves can be created in the future, perhaps when humans have migrated to other planets.
Studies suggest that workers who have specialized training or post-high-school education and those that are younger are most likely to find new work once displaced by AI; others fall into poverty. Job automation has already added to labor-market inequality since at least the 1980s.
This brings to mind dark, Ready Player One-esque images of unemployed masses living in giant, walled-off ghettos. Or, a utopian alternative: A world in which people spend less time working grueling, menial jobs and more time on fulfilling work and leisure, and in which all of our basic needs are met. (In both worlds, AI outsmarts its programmers, gaining some level of sentience and could, in theory, organize against the boss.)
Harry Holzer, an economics professor at Georgetown University who studies the subject, projects that a greater dependence on AI will occur but that the shift will be gradual. As companies transition to using more robot technology, new jobs will be created to respond to a rising demand from consumers who can now afford cheaper products (prices will drop due to employers’ lower production costs).
“The people hurt are not just the people directly displaced by the technology. It’s also the people who in some sense have to compete with the technology or with globalization,” Holzer told The Indypendent. “There are ways in which people can adjust. When technology is implemented, it does some of the tasks that workers have done, not necessarily all the tasks. Now, workers who only did one task on the job — you know, on an assembly line, they tighten some bolts as the car is passing — robots are gonna replace them. But if they have multiple tasks, and the automation does some but not all, there’s a judgment call on the employer’s part: Should he hold on to that worker and train them to do a new task?”
It is unlikely that new jobs will be created at the same rate that old ones are lost, or that enough new jobs will ever be created. And no matter what happens, massive layoffs will occur.
“Going forward, what makes all this more uncertain is that AI seems to have a much wider range of capabilities, even if it doesn’t exist today,” says Holzer.