But exactly which workers are most likely to experience job loss or reduced income when new technologies arrive?
Bryan Seegmillerassistant professor of economics at Kellogg, along with the Kellogg professor of economics Dimitris Papanikolaou and their colleagues, tried to understand better which types of workers have historically been vulnerable to becoming obsolete by technology and how technology-induced career disruptions affected their future earnings. They developed an innovative way to measure workers’ exposure to emerging technology by identifying similarities between the tasks associated with different occupations and the descriptions in new patents. This allowed them to track how emerging technologies affected the exposure of workers in related occupations over time.
As you might expect, they found that manual workers had the highest exposure to emerging technologies, especially from 1850 to 1970. But other patterns were more surprising. In the 1970s, occupations in which people performed routine “cognitive” tasks, such as clerks, technicians, and programmers, also began to experience much greater exposures to technology. And when new inventions appeared, the highest-paid workers in the affected occupations—that is, those with the most advanced skills—saw the biggest slowdown in their wages.
“The most skilled workers have the most to lose,” says Seegmiller. They tend to be “hit harder in terms of their income.”
Winners and losers
In general, technology improves productivity and living standards. But profits and losses are not equally distributed. Each advance can help everyone on average, “but there can be a very specific subset of people who are just absolutely blown away by it,” says Seegmiller.
To better understand which workers have been affected by technological progress historically, Seegmiller and Papanikolaou, along with Leonid Kogan and Lawrence Schmidt at the MIT Sloan School of Management, devised a new way to measure how people’s exposure to technology—that is, their risk of being displaced by new inventions—has changed over time.
The researchers gathered descriptions of jobs performed in more than 13,000 job types from the Dictionary of Occupation Titles database. They then developed an algorithm using tools from natural language processing to compare descriptions of works with patent text from 1840 to 2010, focusing on breakthrough developments. Based on text similarities, the team was able to identify patents that were highly related to work related to specific occupations.
For example, the algorithm matched a 19th-century patent for a knitting machine with occupations such as textile workers and sewers. A patent for a financial account management system combined with financial managers, credit analysts, accountants, bookkeepers, etc.
A college degree won’t help
The team then looked at four broad job categories.
One category was manual occupations, such as electricians and machine operators. Another was interpersonal tasks that required social perception, or the ability to understand and communicate with other people. they included professors and psychologists. Common cognitive tasks involved repetitively performing tasks that typically followed a set list of instructions—for example, clerks and technicians. And non-traditional cognitive occupations required skills such as creative thinking, analyzing information, or leading team members. engineers, surgeons and managers fall into this category.
As might be expected, manual physical tasks were the most exposed to technological change. But the cognitive professions were not immune from danger. Routine cognitive tasks, in particular, began to become much more exposed from the 1970s, as information technology began to take off.
An example was order clerks, whose duties included taking customer orders over the phone, coordinating shipments, and checking order details. In the late 1990s, their exposure to technology increased dramatically. Around this time, several patents were filed for related software, such as an electronic order entry system.
The exposure of workers with a college degree has also increased in recent decades. In the early 2000s, it was almost on par with that of workers without a college degree. “Technologies are breaking into areas they never did before,” says Seegmiller. For example, exposures of various engineering occupations increased in the 1990s due to the introduction of new software and other information technologies that changed the skills required and even automated some of the tasks performed by these occupations.”
And this increased exposure presented a tangible risk to all categories of workers. Based on US census surveys from 1910 to 2010, the team found that increasing exposure to technology was associated with declining employment. And wage data beginning in the 1980s suggests that greater exposure led to lower income. For example, wages for order clerks fell 20 percent relative to other clerk occupations from 1997 to 2010, a time period that saw the rise of e-commerce, which fundamentally changed the profession.
The team then probed deeper to see if there were differences in the harms experienced by different types of workers at a given level of occupational exposure.
For example, the researchers compared workers aged 45 to 55 with workers aged 25 to 35. When faced with the same exposure to technology, in the same type of work, older workers’ wages grew 1.8 times slower over a five-year period. This may be partly because younger workers have invested less time in now-obsolete skills and have more time in the workforce to acquire new ones.
Again, college-educated workers did not fare much better than high school graduates. For both types of workers, the slowdown in income in response to technological progress was similar. “Having a college degree doesn’t necessarily set you apart,” says Seegmiller.
One of the most striking findings came when the team looked at workers who had reached the highest level of income in an exposed occupation—for example, clerks or machine operators who earned relatively high wages compared to their peers. These workers saw their wages slow more than twice as much as the average worker in the same occupation with the same level of exposure to technology. “For people who are really capable, they have a lot of room to fall,” he says.
This pattern was even stronger among high-wage workers in occupations that required a long history of specific types of experience, such as skilled trades such as tool makers, machinists, and electrical equipment repairers. For these workers, “you’re really deep in your investment in those particular skills,” he says.
These wage trends indicated that something more than automation was happening. In the automation scenario, “technology comes along and a robot does what you used to do,” says Seegmiller. But a second type of displacement was also possible: rather than directly replacing workers, technology might change the way their jobs were performed and require people to acquire new skills.
For example, an employee who was very proficient in using a particular record keeping system may need to learn new software, or an experienced machine operator may be faced with unfamiliar equipment. People who had invested a lot of time and effort to master the now obsolete methods could be fired. or if they stayed in their jobs, their wages could stagnate or decline.
“If something new comes along and you’re very good at the old way of doing things, that can be just as difficult for you as a robot coming to replace assembly line workers,” he says.
Learning for Life
Researchers have identified some bright spots. Jobs in the interpersonal category had consistently low exposure to technological change. “One thing that technology can’t do, that has never been able to be replicated, is human-to-human interaction,” says Seegmiller.
And workers who specialized intensively in these interpersonal skills fared better. Even as their exposure to technology rose, their income did not fall as much as in other types of occupations.
Technology has also not been a uniformly negative force. The team conducted a separate analysis to identify patents in various disciplines that did not overlap with professional duties. Exposure to these advances was actually associated with increased workers’ incomes, likely because the inventions had helped them become more productive.
“Not all technology is bad for workers,” says Seegmiller. “But technology hurts certain people.”
So what should workers do to protect themselves from tomorrow’s technologies?
In addition to cultivating interpersonal skills, “it’s really important to be willing to learn and adapt all the time,” she says. Many free or inexpensive online courses can help workers acquire new skills. Policymakers could also develop programs to subsidize training for employees who may soon be displaced.
Moreover, the risk of future technological exposure should not necessarily deter people from pursuing a profession that is valued today. For example, an emerging concern—which was not addressed in this study—is that artificial intelligence will take on complex tasks such as data analysis. This may mean that data analysts will see slower salary growth in the future, but will still be paid relatively high salaries compared to many other occupations that are more isolated from technology. And if these analysts enjoy their work, the rewards of having a fulfilling job could be worth the income risk.
“The thought that ‘AI is going to take over everything, and therefore I should avoid investing in technical skills and becoming, say, a baker’—that’s just too pessimistic,” says Seegmiller.