White-collar jobs are increasingly being automated

Research suggests artificial intelligence and machine-learning solutions are automating white-collar professions at a high rate, though some jobs are more easily automated than others.

Professionals with college degrees and specialized training once felt relatively safe from automation. But recent research indicates that white-collar jobs are increasingly being taken over by machines — maybe not by robots directly, but certainly facing pressures from artificial intelligence (AI) and machine-learning solutions.

A 2019 Wells Fargo research report predicted that automation would cut around 200,000 jobs from the banking industry by 2019. “The hyper-low interest rate environment of the past decade-plus has squeezed banks’ margins to exceptionally tight levels, and given their biggest expense categories is employees, it is no surprise that they are looking for ways to reduce those costs,” notes a Nasdaq report.

The COVID-19 pandemic likely accelerated up the shift toward increased levels of automation. Sales of automation software are expected to rise by 20% this year, after increasing by 12% last year, according to the research firm Gartner. And the consulting firm McKinsey, which predicted before the pandemic that 37 million U.S. workers would be displaced by automation by 2030, recently increased its projection to 45 million.

Further, according to a recent World Economic Forum report, 80% of business leaders are accelerating the automation of their services in response to the pandemic, with 43% expecting new technologies to reduce their workforce. Nearly eight in 10 corporate executives surveyed by Deloitte last year said they had implemented some form of robotic process automation. Another 16% said they planned to do so within three years.

Still, not all jobs will be so easily automated. A new report on U.S. employment from CommercialCafe looked at task-automation compatibility across white-collar occupations and homed in on the white-collar jobs in the U.S. are least compatible with task automation. What the report found is that 60% of U.S. white-collar jobs — which are 86% of all U.S. jobs — are less than 50% compatible with automation. The least-automatable white-collar U.S. jobs are in education, health care, counseling, arts, science, and engineering occupations.

In Wisconsin, white-collar jobs accounted for 32% of total employment in 2020, and 14% of that was in occupations that are at least 50% compatible with task automation. The five office-using occupations that were most compatible with task automation in Wisconsin in 2020 were library technician; agricultural and food science technician; tax preparer; cargo and freight agent; and procurement clerk, each with 97% compatibility with task automation. Meanwhile, the least-compatible occupations in the state were secondary school career/technical education teacher (0%); postsecondary nursing instructor and teacher (0%); lodging manager (1%); postsecondary art, drama, and music teacher (1%); and clergy (2%).

“Automating menial and tedious tasks frees up workers’ resources, which can then be better invested into more complex activities,” notes Ioana Gînsac, a writer for several Yardi product publications, including CommercialCafe. “In this way, automation can be a much-needed helping hand in any industry. According to U.S. Bureau of Labor Statistics (BLS) data, the white-collar occupations included in this analysis accounted for roughly 35% of the nearly 140 million jobs in the U.S. in 2020. Of that share, about 19.4 million jobs were estimated to be at least 50% compatible with automation. In other words, at least 50% of the tasks pertaining to each of these occupations could be automated to some extent. Per the most recent employment data, this represents 40% of white-collar jobs and just 14% of all U.S. jobs.”

Gînsac points out that the automation discourse primarily refers to the automation of tasks and not the entirety of a job. “Almost all existing occupations consist of a complex mix of interdependent tasks and interactions. As such, while it is possible to theoretically generalize that occupations consisting mainly of routine repetitive tasks that follow set procedures can easily be performed by sophisticated algorithms, it is essential to note that very few jobs consist entirely of such tasks.”

For the scope of the study, CommercialCafe deemed an occupation to be “white collar” if the characteristic activity happens at least partly in an office setting — be it virtual or in-person.

U.S. occupations that are least compatible with task automation

The lowest task-automation potential — when none (or close to none) of the aspects of the job are automatable — is mostly associated with occupations that are also essential to the functional fabric of society, notes Gînsac. High-complexity jobs in arts, education, health care, counseling, social work, and earth science and engineering are less than 5% compatible with task automation. Jobs in these occupations represent nearly 17% of white-collar employment and only 5.35% of all U.S. jobs.

“At 0% task-automation potential, we found secondary school career/technical education teachers, along with postsecondary nursing instructors and teachers,” says Gînsac. “The former accounted for 73,530 jobs last year, which translates into 0.17% of white-collar jobs and 0.05% of all U.S. employment. Similarly, there were about 61,100 jobs filled for nursing instructors and teachers in 2020, accounting for 0.14% of all white-collar employment and 0.04% of all jobs in the country.

“Postsecondary art, drama and music teachers (1% automation compatibility) added up to 91,170 jobs in 2020, which accounted for 0.21% of white-collar jobs and 0.07% of total employment,” Gînsac adds. “There were also 31,790 lodging managers across the U.S. last year, making this the smallest cohort among the five occupations highlighted here. Also coming in at 1% task-automation compatibility, lodging managers accounted for 0.07% of white-collar jobs in the U.S. last year, which equated to 0.02% of all employment in the country.

“Conversely, preschool teachers (except special education) made up the largest group of the five lowest-compatibility occupations,” Gînsac notes. “In 2020, there were 370,940 jobs filled for this occupation in the U.S., which represented 0.84% of white-collar employment and 0.27% of all U.S. jobs. According to the WillRobotsTakeMyJob score, roughly 2% of tasks associated with this occupation would be compatible with automation.”

The other eight occupations to be estimated at 2% task-automation compatibility include clergy, conservation and soil & plant scientists, special education teachers, and counselors. These occupations account for a combined total of 1.91% of white-collar jobs and 0.6% of all U.S. employment.

U.S. occupations most compatible with task automation

On the opposite end of the spectrum lie the occupations in which more than 95% of associated tasks could be automated, Gînsac explains. Often, these are aspects such as clerical and administrative tasks, as well as sampling, measurements, and logistics management. The five occupations that topped this list accounted for a combined total of roughly 0.24% of total U.S. employment: Procurement clerks; library technicians; tax preparers; cargo and freight agents; and agricultural and food science technicians each scored a nearly perfect 10 for automation compatibility. Roughly 97% of tasks associated with each of these five occupations could be computerized.

“Although the dispersion of these jobs across the U.S. will vary, what we can look at is how much each of them represents within the total employment,” says Gînsac. “For instance, last year, roughly 96,510 cargo and freight agents were employed across the country, comprising 0.22% of white-collar jobs and 0.07% of the total U.S. employment. Second in line by number of jobs were library technicians: BLS data showed 89,070 jobs filled for this occupation in 2020, which amounted to 0.20% of white-collar employees and 0.06% of all U.S. workers.

“Next, tax preparers numbered around 62,600 last year, rounding out to 0.14% of U.S. white-collar jobs and 0.05% of total employment,” she adds. “At the same time, procurement clerks working in the U.S. accounted for 0.14% of white-collar employment and 0.04% of total jobs in the country. Finally, in 2020, there were 21,940 agricultural and food science technicians working across the U.S., which represented 0.05% of white-collar jobs and 0.02% of all jobs in the country.”

Ultimately, automation of jobs, be they blue collar or white collar, shouldn’t cause too much concern. “We see there are a small number of occupations where a large number of tasks are suitable for machine learning,” Martin Fleming, chief economist at IBM, told CNBC Make It back in 2019. “But there are a large number of occupations where a small number of tasks are suitable for machine learning.”

In short, “the thought that robots are stealing our jobs is nonsense,” he said.

Click here to sign up for the free IB Ezine — your twice-weekly resource for local business news, analysis, voices, and the names you need to know. If you are not already a subscriber to In Business magazine, be sure to sign up for our monthly print edition here.