Richard Florida is a co-founder and editor at large of CityLab and a senior editor at The Atlantic. He is a university professor in the University of Toronto’s School of Cities and Rotman School of Management, and a distinguished fellow at New York University’s Schack Institute of Real Estate and visiting fellow at Florida International University.
A new study uses artificial intelligence to find that jobs done by highly skilled workers are the most likely to be affected by AI.
When most people think of the connection between technology and jobs, they think of robots and automation taking over relatively unskilled jobs like factory work. And thus, the biggest toll from these technological advances would be on already hard-hit manufacturing regions of the Rust Belt. But a new wave of developments in artificial intelligence may have a greater effect on high-skilled jobs and high-tech knowledge regions.
That’s the key takeaway from a new study out today from the Brookings Institution. The study by Mark Muro, Jacob Whiton, and Robert Maxim takes a close look at the potential of artificial intelligence—or AI—to automate tasks that until now have required human intelligence and decision-making. As they put it: “Unlike robotics (associated with the factory floor) and computers (associated with routine office activities), AI has a distinctly white-collar bent.”
The Brookings study bases its analysis on a set of “exposure scores,” developed by Michael Webb, a doctoral student at Stanford University, which essentially gauge the potential effects of AI on different jobs. In fact, Webb uses AI to study AI, using machine learning to search all U.S. patents to identify the capabilities of AI, and to connect that data to jobs and tasks that could be taken over by AI technology—tasks like certain medical diagnoses that doctors perform today. Brookings, in turn, uses those scores to assess how AI will affect occupations and places. In doing that, Brookings’ analysis quantifies degree of potential exposure but not whether it will be positive or negative.
What does the Brookings study find? First, while A.I. will likely affect a wide array of work and jobs, its largest effects will be confined to a much smaller segment of jobs. Overall, AI will, in some way, influence more than 95 percent of jobs. As the study notes: “Fully 740 out of the 769 occupational descriptions Michael Webb analyzed contain a capability pair match with AI patent language, meaning at least one or more of its tasks could potentially be exposed to, complemented by, or completed by AI.”
But, as the chart below shows, less than a fifth (just under 18 percent) of U.S. jobs, 25 million or so, are threatened by high exposure to AI. Roughly a third (34 percent or 48 million jobs) face a medium level of exposure; and a little fewer than half (48 percent or 67 million jobs) face low or no exposure to AI.
Share of jobs by AI exposure, 2017
But, AI is different from automation or robots in that it is more likely to affect higher-skilled work. This can be seen in the chart below, which shows that while AI is likely to affect manufacturing and agricultural work, it is much more likely than robotics or automation to affect higher-wage, higher-skill occupations done by college graduates, and people with advanced or professional degrees.
Average standardized AI exposure by education level, 2017
The next chart drills down further into the more fine-grained categories of jobs that will likely be affected most by AI. A number of lower-skilled occupations rank highly, like farming, manufacturing, mining, and construction. But also exposed are high-skill jobs like professional, scientific and technical services; information; and finance and insurance.
“High-tech digital services such as software publishing and computer system design—that before had low automation susceptibility—exhibit quite high exposure, as AI tools and applications pervade the technology sector,” the study points out. The jobs that are least exposed include educational services and arts and entertainment, alongside lower-skilled jobs in retail and accommodation and food service, that are personal services.
AI is also more likely to affect male, white, and Asian-American workers, because of their over-representation in professional and technical occupations, as well as prime-age workers (25-64), according to the study.
Average standardized AI exposure by sex, age, and race-ethnicity; 2017
AI is likely to hit hardest at a combination of leading tech hubs and older manufacturing regions. San Jose—the heart of Silicon Valley—tops the list of metros that are most exposed to AI. Seattle is fifth; Salt Lake City is eighth; Ogden, 10th; and Durham in the North Carolina Research Triangle, 12th.
Smaller high-tech hubs like Boulder and Huntsville, Alabama, are also highly exposed. Manufacturing metros like Detroit; Grand Rapids; Louisville; and Greensboro-High Point, North Carolina, face a high level of exposure, as well as smaller manufacturing centers like Elkhart-Goshen, Indiana; and Dalton, Georgia. And the Sun Belt metros Nashville, Atlanta, and Charlotte have high levels of AI exposure due to the significant presence of management and finance occupations, as well as some manufacturing.
Service economy and recreational metros—both large ones like Las Vegas; Cape Coral-Fort Meyers and Deltona-Daytona Beach, Florida; and smaller ones like Hilton Head and Myrtle Beach, South Carolina; Ocean City, New Jersey; and El Paso and McAllen, Texas, have among the lowest levels of AI exposure. AI is significantly less of a threat to smaller and more rural places than other forms of automation and robots, the study notes.
Top 15 and bottom five metro areas and NECTAs by average standardized AI exposure, 2017
History shows us the introduction of new labor-replacing technology does not occur in a vacuum. Not only is it typically associated with increasing worker anxiety, but also with a potent political backlash. In the early 19th century, the introduction of machinery in British factories fueled the Luddite revolution. The last wave of robotics and automation technology hit hardest at manufacturing jobs and regions, helping to fuel the populist backlash that elected Donald Trump.
The coming widespread use of AI could extend the kind of fear and anxiety felt by lower-skilled manufacturing workers and regions to more affluent and educated professional and technical workers living in many leading tech hubs. These workers and places have, to date, largely been spared by the previous wave of automation and robotics. Might an even larger political earthquake be in the offing?