PwC says AI is raising the bar for entry-level jobs

Artificial intelligence is changing what employers ask of new workers, according to PwC, and the shift may be making many entry-level jobs harder for recent graduates to land. PwC’s 2026 AI Jobs Barometer says roles at the bottom of the career ladder increasingly require judgment, leadership and other skills that used to develop later on the job.

The firm’s analysis, based on more than 1 billion job postings, found that entry-level jobs in occupations highly exposed to AI are seven times more likely to require skills historically tied to more senior workers. PwC said 52% of new skills appearing in entry-level postings in the most AI-exposed occupations fit that senior-skills category, compared with 7% in the least AI-exposed occupations.

PwC calls the pattern “seniorization.” According to the firm, postings for these redesigned entry-level roles have risen 35% since 2019, while traditional entry-level openings fell 10% over the same period.

First jobs are asking for more

Dan Priest, PwC’s U.S. chief AI officer, told Fortune that employers are changing their expectations for early-career roles as AI takes over more routine work. He said companies are putting more weight on human capabilities such as judgment, communication, creativity, collaboration and leadership.

Priest cautioned Fortune against portraying the shift as employers using AI as an excuse to cut off young workers. He said the evidence shows a change in what companies want from entry-level hires, and that employers, educators and policymakers need to help workers build those abilities earlier.

The findings line up with other labor-market signals cited by Fortune. A Harvard working paper that studied 62 million workers found junior hiring fell nearly 8% within six quarters at companies that adopted AI, with the change showing up as fewer new roles rather than layoffs. The New York Fed reported recent graduate unemployment at 5.7% in the fourth quarter of 2025, while recent graduate underemployment stood at 42.5%.

PwC’s data also shows that early-career hiring has grown overall. Priest told Fortune that PwC’s global early-career dataset counted about 11 million early-career job postings in 2025, up from 7.3 million in 2018 and 3.2 million in 2012. In highly AI-exposed occupations, however, he said growth is increasingly concentrated in roles that call for more judgment-heavy work.

Productivity gains are uneven

PwC’s barometer also found stronger productivity growth in AI-exposed sectors. The firm said companies in the most AI-exposed sectors recorded 34% labor productivity growth since 2018, compared with 24% in the least exposed sectors.

Among the top-performing 20% of the most AI-exposed companies, PwC reported average labor productivity growth of 163% since 2018. Priest told Fortune that companies getting more value from AI are changing workflows and decisions rather than only adding new tools.

The report does not support a simple conclusion that AI-heavy companies are cutting overall headcount. PwC said headcount at AI-heavy companies is growing faster than at less-exposed peers. The harder question, raised by the firm’s entry-level findings, is which workers are getting those jobs.

Growth is stronger outside AI-exposed work

PwC found that job-posting growth since 2012 has been faster in occupations with less AI exposure. By 2025, the lowest AI-exposure quartile had 4.7 postings for every posting in 2012, while the highest exposure quartile had 1.9.

Fortune reported that the faster-growing roles include construction, plumbing, welding, kitchen staff, agricultural workers and health care aides. These jobs involve physical, location-based or human-facing work that current AI systems do not directly replace.

Priest told Fortune that the barometer tracks employer demand through job postings, not every aspect of worker supply or hiring outcomes. He said companies should redesign routes into work if AI changes the first step on the career ladder.

This story draws on original reporting from Fortune.