At the start of the first Terminator movie, Sarah Connor, unknowingly the future mother of Earth’s resistance movement, is working as a waitress when Arnold Schwarzenegger’s Cyberdyne Systems Model 101 Terminator is sent back through time to kill her. But what if, instead of attempting to murder her, Skynet’s android assassin instead approached the owner of Big Jeff’s family restaurant, where Sarah worked, and offered to do her shifts for lower wages, while working faster and making fewer mistakes? The newly jobless Sarah, unable to support herself, drops out of college and decides that maybe starting a family in this economic climate just isn’t smart. Hey, presto: No more John Connor.
This, in a somewhat cyberbolic nutshell, is the biggest immediate threat many fear when it comes to automation: Not a robopocalypse brought on by superintelligence, but rather one that ushers in an age of technological unemployment.
Some very smart people have been sounding the alarm for years. A 2013 study carried out by the Oxford Martin School suggested that some 47% of jobs in the U.S. could be automated within the next two decades — only 12 years of which now remain following the publishing of the study. Like the old Hemingway quote about bankruptcy, that it happens “gradually, then suddenly,” the authors suggested that, in the first wave, office and admin support workers, as well as laborers in production occupations, would be Thanos-snapped out of existence. In the second wave, every task involving finger dexterity, feedback, observation and working in confined spaces would be eaten up by software.
Why are there still so many jobs?
To date, this hasn’t happened in such catastrophically large numbers. In fact, the hiring carried out by big tech firms that have most heavily invested in automation has outpaced many other industries. Amazon, which once laid off its human editors in favor of algorithmic recommender systems and is constantly working to roboticize its warehouses, hired an extra 175,000 people as the coronavirus lockdown commenced in March of last year. Other tech companies like Netflix have also not slowed down on hiring, even at a time when COVID was cratering many industries.
These companies have, of course, benefitted from a very trying time in global history. Streaming media companies, communication companies like Zoom, device makers like Apple, and e-commerce “everything stores” like Amazon were perfectly placed to benefit from the world being stuck at home. But it illustrates the complexity of the situation. Potentially human-replacing A.I., robotics, and assorted technological infrastructure made these “unicorns” unicornier, which, in turn, meant that they could hire more people.
These effects can seem counterintuitive. In an essay titled “ Why Are There Still So Many Jobs? The History and Future of Workplace Automation ,” Massachusetts Institute of Technology (MIT) professor of economics David Autor looked at the quadrupling of ATMs between 1995 and 2010, and how it impacted the number of bank tellers employed in banks. An ATM isn’t, of course, the same as an advanced robot, but you could sensibly assume that an extra 300,000 money-dispensing ATMs would lead to a decline in the number of people hired to dispense money.
In fact, bank teller employment actually rose by 50,000 during this same period. ATMs meant that more bank workers could be freed up to focus on what Autor calls “relationship banking.” Autor notes that technology means bank workers are no longer primarily “checkout clerks, but … salespersons, forging relationships with customers and introducing them to additional bank services like credit cards, loans, and investment products.”
Augmenting, not replacing
This is the big promise of tools like A.I. — that they won’t so much replace humans as they will augment humans. They will, we are told, take away the dull, dirty, and dangerous jobs, while allowing humans to focus on bigger value-adding tasks. If this is accurate, that’s great. No one is bemoaning the fact that technology (and civilization in general) did away with child chimney sweeps in Victorian England. Perhaps we wouldn’t be too sad about certain mindless data-angry tasks or dirty jobs — not to mention potentially fatal ones — being snapped up by robots.
A growing number of jobs are likely to be hybrid jobs in which humans work alongside machines.
The makeup of jobs changes over time. In 1800, 90% of the U.S. lived and worked on farms. Today, the overwhelming majority live and work in cities. A recent MIT study, “ Work of the Future ,” noted that 63% of today’s jobs did not exist just 80 years ago in the early 1940s. Since 1990, over 1,500 new occupation roles have appeared as official job categories, including software engineers, SEO experts, and database administrators. Many of these are technological, but other job types are “high-touch” roles built around personal interaction that are seemingly only becoming more important as our lives become more digital.
A growing number of jobs are likely to be hybrid jobs in which humans work alongside machines. In some cases, these will be technologies such as RPA (robotic process automation) tools, which can sit on the desktop of human workers and provide them with pointers on how to do their job better, such as prioritizing tasks or remaining compliant in compliance-heavy occupations. Emotion-sniffing A.I. can help identify the emotions of callers and route them through to the correct human operator at a call center.
Meanwhile, tech companies benefit from what is called AAI ,or “artificial artificial intelligence,” in which humans help perform tasks A.I. is currently incapable of. Twitter, for instance, employs human contract workers, called judges, whose job it is to interpret the meaning of different search termsthat trend on the service. Meanwhile, in Amazon fulfillment centers, robots such as those made by the Boston-based Kiva Systems (purchased by Amazon almost a decade ago) are used to ferry around racks of shelves, taking them to the human “picker,” who can then use their fine motor control to pick the right item up for packing.
The future of employment
The big question is what this will all mean for human employment going forward. As artificial intelligence gets smarter, more tasks that currently require humans can be automated. There are A.I. bots able to perform certain tasks once deemed worthy of a high social capital job like a lawyer. Right now, human drivers can be hired to oversee the A.I. driving autonomous vehicles, thereby providing those people better, more sociable working hours during which they don’t have to be on the road for days at a time.
But will they always? Probably not. The same goes for picking objects in Amazon warehouses and, potentially one day, delivering the packages through customer’s doors. However, as some of these low-hanging fruits are plucked, humans will be able to take on the higher-hanging ones that machines are not yet capable of.
Automation alone isn’t going to steal jobs like some people fear. It’s a far more complex landscape than that simple way of looking at things suggests. It will dominate some jobs, but also usher in new forms of employment, much of it (although not all) related to developing, maintaining, or working with this new technological infrastructure. As shown by tech giants’ hiring sprees, and Autor’s research into ATMs, the idea that companies investing in tech is necessarily bad for human workers isn’t a foregone conclusion.
What technology may — and probably will — do is to exacerbate societal trends. Around four decades ago, many U.S. workers experienced a divergence on the trajectory of wage increases and productivity growth. Technology will, as many have pointed out, likely help to hollow out the middle classes, pushing some upwards in both earnings and quality of jobs, while making things tougher for others. It will also mean a society in which participants are constantly retraining and upskilling, partly to stay on the right side of the technological replacement wave. But there is far more nuance to this picture than is sometimes presented.
The American history Melvin Kranzberg famously said that “technology is neither good nor bad; nor is it neutral.” The same can be said for its likely impact on the job market. It’s complex. But where there is chaos, there is also opportunity.
