|
Getting your Trinity Audio player ready...
|
For the past few years, the dominant fear around artificial intelligence and employment has centered on mass layoffs, robots replacing workers wholesale, and entire job categories disappearing overnight. The data emerging in 2026 tells a more subtle and, in some ways, more troubling story. AI is not showing up primarily as pink slips. It is showing up as smaller raises, slower wage growth, and a widening gap between workers who can use the technology effectively and those who cannot.
Research from Apollo Global Management, one of the most closely watched studies on this trend, found that workers in occupations with high exposure to AI saw real wage growth come in 6.7 percentage points slower than comparable workers in less-exposed roles after 2023, with no detectable increase in job losses tied to the technology. In other words, people are largely keeping their jobs, but the pay increases that would normally accompany productivity gains and inflation adjustments are not materializing at the same pace. Apollo’s analysis found the effect was not evenly distributed. The bottom wage quartile absorbed a 10.7 percent penalty, the second quartile saw 5.4 percent, and the third quartile experienced a 4.0 percent hit, while top earners and workers in physically demanding blue-collar roles showed no measurable effect at all. Service occupations, including childcare workers, servers and social workers, saw the steepest relative wage decline at 24.3 percent, though Apollo cautioned that sample size for that category was smaller than for other groups studied.
The scale of who this affects remains relatively contained for now. Apollo’s research puts the number of workers currently experiencing this wage compression at roughly 5.8 million people, or about 3.7 percent of the U.S. labor force, with an estimated $28 billion in aggregate lost labor income per year. The Bureau of Labor Statistics projects that the population of workers in high-exposure occupations could grow to 5.9 million by 2032, suggesting the phenomenon is more likely to widen gradually than to trigger a sudden shock.
Separate research reinforces the pattern from a different angle. A study cited by Forbes found that starting wages at companies heavily exposed to AI fell 4.5 percent following the launch of ChatGPT, with junior and mid-level pay declining even as senior compensation held steady. That divergence points to a labor market increasingly split along experience lines, where AI tools are capable enough to compress the value of entry-level and early-career work while leaving senior roles, which typically involve judgment, oversight and accountability, comparatively untouched.
Entry-level workers appear to be absorbing a disproportionate share of the disruption more broadly. Research from Stanford found relative employment for workers aged 22 to 25 in highly AI-exposed occupations fell between 13 and 16 percent, while separate UK government data found job postings in exposed roles down 38 percent. A 2026 employer survey from ZipRecruiter found that nearly four in ten companies have already shifted basic data entry and processing tasks away from entry-level employees and onto AI systems, even as overall hiring at those same companies continued to grow. The survey also found a significant support gap behind that shift, with only 22 percent of employers offering mandatory AI training to all staff, leaving many early-career workers to build AI fluency largely on their own initiative.
At the same time, workers who have developed strong AI skills are seeing the opposite effect. PwC’s 2026 Global AI Jobs Barometer, which analyzed more than a billion job advertisements across dozens of countries, found the average wage premium for workers with AI skills climbed to 62 percent, up from 57 percent the year before and just 25 percent two years earlier. That premium varies significantly by industry, reaching as high as 118 percent in consumer markets while sitting closer to 16 percent in government and public sector work. Jobs specifically requiring AI skills are growing roughly 69 percent year over year, nearly eight times faster than the overall job market’s 9 percent growth rate, according to the same report.
PwC’s research also distinguishes between two different ways AI is reshaping roles. In what the firm calls professionalised roles, AI takes over routine tasks while leaving employees responsible for judgment, context and accountability, and pay for these positions rose 37 percent since 2021. In democratised roles, where AI reduces the level of expertise required to do the job, pay grew more slowly, at 26 percent over the same period. The distinction suggests that whether AI helps or hurts a worker’s paycheck depends heavily on whether the technology elevates their remaining responsibilities or simply narrows what expertise is needed to do the work at all.
Not every analysis frames the moment as a distinct AI-driven shock. A separate review from Futurum Group argued that much of the broader hiring softness seen across the U.S. and UK in 2026 reflects a cyclical slowdown rather than automation displacing workers economy-wide, with the clearest AI effect concentrated narrowly in entry-level software and customer support roles where tasks are easy to verify and automate. That more measured read does not contradict the wage compression data so much as it narrows where the effect is actually concentrated, suggesting the story is less about a sweeping AI jobs apocalypse and more about a specific, identifiable slice of the workforce absorbing real financial pressure while headline job numbers stay largely intact.
Taken together, the emerging picture suggests AI’s early economic impact on American workers looks less like the dramatic job-loss scenarios that dominated public discussion a few years ago and more like a quieter redistribution of pay, with entry-level and lower-wage workers absorbing measurable losses while workers who can demonstrate AI fluency capture outsized gains. Whether that pattern widens or stabilizes over the next several years will likely depend on how quickly companies invest in structured AI training rather than leaving workers to adapt on their own. Continuing coverage of how AI is reshaping labor markets and business strategy is available on Business Tech. Additional detail on the wage compression findings is available through Apollo Global Management’s research, and PwC’s full 2026 Global AI Jobs Barometer report can be found on the firm’s official site.
