We’ve all heard the debate about AI and jobs: an apocalypse is coming, only 18 months left to save employees, no job will remain the same. Former The White House AI czar, David Sacksshortly before he resigned amid a row over policy, argued that the doomsday predictions of figures like Dario Amodei and Sam Altman were a “Damage to public trust.” Amodei and Altman have recently scaled back their predictions, Assets was one of the first to noticebut fear and anxiety remain among Generation Z job seekers.
There are now real cracks in the entry-level career ladder, experts like Stanford’s Erik Brynjolfsson argues for clear signs in the data of disruption in AI-at-risk jobs, even if the broader macro picture has shown that the earthquake is not yet here.
A new PwC analysis of more than a billion job postings shows a more precise and, for young workers, more worrying shift: AI is not displacing entry-level jobs. This makes it something that beginners can’t get.
The AI job barometer 2026The study, released Monday, finds that entry-level positions in high-AI occupations are now seven times more likely to require skills that were historically used later in a worker’s career – things like strategic decision-making, stakeholder management, leadership and judgment. In the occupations most impacted by AI, 52% of new skills appearing in entry-level job postings were skills traditionally associated with experienced workers. In the jobs least exposed to AI, this figure was 7%.
The seniorization of entry-level work
PwC calls this “seniorization,” and the numbers are dramatic. Job postings for these reimagined entry-level positions — those that now require a 22-year-old to demonstrate the skills of a 35-year-old — have increased 35% since 2019. Traditional job openings for entry-level positions have declined by 10% over the same period.
This is the mechanism behind a labor market anomaly Assets has followed the last year. A Harvard working paper The company, which analyzed 62 million workers, found that junior hiring at companies adopting AI fell nearly 8% over six quarters—not through layoffs, but through a quiet freeze on new jobs. The unemployment rate among college graduates reached 5.7% in the fourth quarter of 2025. according to the New York Fedabove the national rate and a near reversal of the historical norm. Current underemployment of college graduates is 42.5%.
The PwC data offers an explanation. Entry-level positions haven’t disappeared, but the job description has been quietly promoted up the skills ladder without notifying the people trying to get their foot in the door.
Dan PriestPwC’s U.S. chief AI officer was careful not to portray this as manipulation of the system by employers.
“I would be cautious about portraying it as employers using AI as an excuse for anything,” he said Assets. “What it shows is that employers are changing their requirements for entry-level positions.”
He acknowledged the consequence directly: “As entry-level work becomes more demanding, employers, educators and policymakers all have a role to play in helping people build these skills sooner. The answer cannot simply be to raise the bar and hope that talent emerges.”
“The broader story is that AI is changing the shape of entry-level work,” Priest continued. “As AI takes over more routine tasks, employers are placing more value on uniquely human skills and requiring entry-level professionals to bring those skills to bear sooner than in the past.”
In other words, the bar has been raised as AI not only transforms the capabilities of employees, but also the needs of employers. The infrastructure to clear it, not so much.
“The message of education is not simply about teaching more AI,” Priest said, but about teaching AI along with the human skills that make AI useful. “The future advantage lies in people who can guide, challenge and apply AI to real-world problems, rather than just nudging them.”
The productivity boom and its limits
The other side of the PwC barometer complicates the simple story. Companies in the most AI-exposed sectors have seen labor productivity growth of 34% since 2018, compared to 24% in the least exposed sectors. At the top of the distribution is what PwC calls the “superstar effect”: the top 20% of most AI-exposed companies achieved average labor productivity growth of 163% since 2018 – almost five times higher than the average for all AI-exposed companies overall. What is more counterintuitive is that the number of employees in AI-heavy companies is growing faster than at least their least exposed competitors.
This disrupts the fundamental narrative of “AI means fewer workers,” and the disruption is real. In the companies that use it most effectively, AI appears to be expanding what organizations can do, rather than simply replacing the people who used to do it.
“What matters to leaders is that the gap is real,” Priest said. “The companies that are getting greater value from AI aren’t just adding tools. They’re redesigning workflows, rethinking decisions, and integrating AI into the way work gets done.”
But behind the headline lies a compositional question that the barometer cannot fully answer: Hire more, yes – but hire who? The finding of seniorization suggests that the entry level of these new employees is decreasing despite the overall number increasing. The most AI-exposed companies are hiring—and they’re looking for workers who can lead AI, apply judgment, and manage stakeholders. People who have not shown up for their first interview in the past.
Priest clarified that the entry-level market is still growing in absolute terms. In PwC’s global entry-level career data set, about 11 million entry-level positions were advertised in 2025, up from 7.3 million in 2018 and 3.2 million in 2012. However, he acknowledged that growth in heavily AI-exposed occupations is increasingly concentrated in judgment-oriented entry-level positions.
“It’s not about entry-level workers disappearing, it’s about the skills employers are looking for evolving,” he said.
“It’s not just technical AI skills that are becoming more important in AI-at-risk jobs,” he continued. “Employers are increasingly looking for judgment, communication, leadership, creativity and collaboration,” he added – not something that has developed over the course of workers’ careers in the past. This is a paradox for Generation Z job seekers – and the schools and internships that train them.
Where jobs are actually growing
There is another wrinkle in the data that contradicts the usual narrative. Since 2012, job posting growth has been significantly faster in less AI-exposed occupations than in highly AI-exposed occupations. In 2025, the lowest AI exposure quartile was 4.7 posts per post in 2012. The highest exposure quartile: 1.9.
The occupations driving this growth are what you would expect: construction, plumbing, welding, kitchen staff, farm workers, health care workers. This is physical, location-based, human-centered work that current AI cannot directly replace.
There is a version of this insight that is transformed into reassurance: trade is booming, the economy is balancing, but this reading is too convenient. It ignores the wage and status realities of many of these roles, ignores that the workers currently locked out of entry-level positions at risk of AI did not plan to pursue a career as plumbers, and avoids the structural question of who exactly is taking them somewhere.
According to Priest, the barometer is based on job postings and therefore only tells us how employer demand is changing. In his opinion, however, the results make it clear that the “transition” must be consciously designed.
“If AI changes the first step of the career ladder, then companies have a responsibility to redesign the paths to work, not just the work itself,” he said. The most successful Fortune 500 companies are those that invest heavily in transforming the workforce, he added, citing first-hand experience.
Overall, the picture that emerges is an AI economy that is achieving real productivity gains while quietly restructuring who can participate. The entry-level job wasn’t killed. It was funded – and the funding came without notice, without a training program and without the policy framework that could have mitigated the transition.
“AI is different from some previous waves of technology,” Priest said, “because it affects a broader range of professions at the same time.”
It is not just limited to technology but also extends to professional services, finance, healthcare, education, operations and many other areas.