The biggest winners — and losers — in the coming AI job apocalypse
Disruption from artificial intelligence is just one of the major structural shifts that will define the American economy over the next few decades. But while white-collar professionals stare down a turbulent job market and declining wages, a large, unexpected segment of the workforce may be safely on high ground: blue-collar jobs. The likes of skilled trades, intensive manual labor, and jobs that require a combination of physical, knowledge, and social work will not be upended. In fact, as other long-term trends take hold, many of these working-class roles are poised for a job explosion.
As older Americans retire in massive numbers and leave behind open jobs, and as demand for healthcare, green energy, high-tech manufacturing, and construction increases, blue-collar jobs are going to see a substantial boom. ChatGPT will not replace the nurse who tends to you at the hospital or the construction worker who remodels your kitchen. The winners of the AI revolution will be the technicians, nurses, and plumbers who keep the new economy running after the machines have taken over the office.
The new economy could be a boon for the working class
Describing the nature of factory work in the '90s, the American poet Philip Levine wrote, "We stand in the rain / in a long line waiting at Ford Highland Park. For work. / You know what work is." At the time, factory jobs in the Rust Belt had dwindled, job-seeking became undignified, and the nature of work grew monotonous and alienating. The decline of manufacturing sped up precipitously in the early 2000s as globalization pushed jobs overseas, low-skilled work was automated away, and the economy shifted toward desk jobs with the advent of the internet.
Today, there's a hopeful reversal for American blue-collar jobs. While cognitive workers fight over rapidly disappearing remote jobs, industries such as semiconductor manufacturing have the opposite problem — too many jobs to fill. And that's true across working-class industries. Two years of a historically tight labor market and low unemployment have pushed wages up for lower- and middle-income workers, stimulating a working-class job boom that has seen 9.6 million job openings nationwide and what economists are calling "full employment." Over the next decade, the US Bureau of Labor Statistics forecasts that the fastest-growing occupations will be in healthcare, transportation, renewable energy, and high-tech manufacturing.
Nearly half of all job gains will come from the healthcare and social-support sectors. By 2032, one out of six new workers will be a home-health or personal-care aide. Construction, which has added 15,000 jobs a month over the past year, also has room to keep expanding.
While manufacturing jobs as a whole are expected to stay flat, spending in this industry has boomed to $200 billion each year, tripling in the past five years. And certain sectors are expected to grow: The electrical-equipment and semiconductor industry is projected to reach 3% annual growth over the next decade. Chip manufacturing will continue to be a national priority, and that means a lot of well-paid jobs that require only a high-school education.
"There's going to be significant demand in manufacturing for the sorts of non-BA technician jobs," Mark Muro, an expert on technology and workforce development and a policy director at the Brookings Institution, told me. Within chip manufacturing, roughly 50% of entry-level workers have only a high-school or equivalent education level, compared with 38% for all other industries. Over the next 10 years, as the global semiconductor industry is primed to reach $1 trillion, another 50,000 American jobs will be added for roles such as electrical assemblers and industrial technicians.
While not all working-class jobs will be safe from automation — retail workers, for instance, do not have a rosy outlook — there has been improvement across the board for lower-income workers. The initial shock of the pandemic hit low-wage workers the hardest, but the recovery since has been more auspicious. Wages have grown nearly three times as fast as in any other recovery period. The wages of the lowest 10th percentile of the workforce — earning an average of $12.50 an hour — grew only 3.9% from 2009 to 2017. But from March 2020 to March 2022, wages grew by 5.7%. Despite inflation, real wages are higher for blue-collar, nonmanagerial, and lower- and middle-wage workers than they were before the pandemic.
Other economic barometers also show steady improvements that are likely to continue. Over the past decade, the poverty rate in America has fallen from 14.8% to 11.5%. The net worth of the average household in the bottom 50% of Americans is up more than 45% from before the pandemic.
If a recession is averted, the continuation of a tight labor market will continue to drive wages up, further reduce the poverty rate, and boost household income and net worth. Adding to the tailwinds is AI's adoption, which is likely to augment these jobs, leading to higher productivity growth over the coming decade. Increased labor productivity in the construction industry would be a boon for economic growth since housing scarcity remains one of the most pressing issues for Americans.
Who's safe from AI?
In the World Economic Forum's recent joint report with Accenture, 19,000 individual tasks across nearly 900 occupations were analyzed for how much they could be affected by AI. The findings were pretty straightforward: Jobs with a high degree of personal interaction and nonroutine physical tasks faced the least amount of disruption from AI. Other recent projections by the Bureau of Labor Statistics and McKinsey have come to similar conclusions.
"What characterizes the physical labor jobs that are safe for the next five or 10 years are things that are in an unpredictable physical environment," Kweilin Ellingrud, a McKinsey Global Institute director, told me. "If it's in a predictable physical environment, and it's repetitive, then either a robot — likely even before generative AI — or some sort of generative AI can create the automation to do that well."
One influential explanation for this is known as skill-biased technological change, a theory that says routine roles that are susceptible to automation and computerization will see declining wages, while nonroutine roles that require dexterity or human interaction will see increased demand and wages.
That's what recent research has found: Jobs such as air-conditioning installer, teacher, and hairdresser — jobs that are not routine — will be relatively unscathed over the next five to 10 years. Research led by OpenAI found that 4% of workers, including painters, carpenters, and roofers, had zero tasks that could be influenced by AI. And even though technology could infiltrate more of these roles, they fundamentally cannot be replaced by machines. Ellingrud said, "Even though we have Roombas, we still have house cleaners because the Roomba can only do so much and isn't all that effective."
Instead of replacing these jobs, AI will likely benefit specific roles by making it easier to do the most routine parts of the job. Nurses, for example, could spend less time collecting and entering information into a medical-record system. For an industry notorious for high hourly billing rates, allowing nurses to redeploy their time on more critical tasks — such as taking care of patients — would save Americans billions in annual healthcare costs. Similarly, the construction industry could benefit from new technology such as AI drones to do inspections or AI sensors that measure things, freeing up workers to do more.
"You may get a bifurcation within these groups where the hands-on physical work remains, but generative AI picks up and substitutes for a lot of the management supervising," Muro told me. He added: "There are these jobs that are in a middle ground where the physical work may remain but the supervision might be more exposed." But instead of replacing workers, this kind of integration could help these types of roles be more productive, have more demand, and earn more money.
Safe ground
Blue-collar jobs are not only safe from AI's impact but also physically safer with the tech. AI's adoption in the workplace can help improve safety via real-time monitoring of risks and alerts for equipment issues or injuries. Each year in the US, there are 2.7 million workplace injuries and over 5,000 deaths as a result of workplace accidents, so any improvements would be significant.
Though many of these working-class jobs may be physically or emotionally laborious, they are durable for the future. Broadly speaking, the jobs that will come out on top of the AI shuffle are good-quality jobs that provide steady work hours, upward mobility, and career support, as well as "fair compensation and a degree of voice," Muro said.
As the economy gets increasingly precarious because of AI, the winners of this AI revolution will not need to swim a technological triathlon against oncoming waves — they will already be on elevated terrain, safely and skillfully working away.
Emil Skandul is a writer on technology and urban economics, and a Tony Blair Institute fellow.