The two-track labour market: PwC's 2026 AI Jobs Barometer

TL;DR

• PwC's 2026 Global AI Jobs Barometer analyses more than a billion job listings across six continents.
• Roles "professionalised" by AI, where automation of routine tasks raises the premium on human judgement, are growing about twice as fast as roles "democratised" by AI, with 42 percent faster salary growth.
• The wage premium for workers with AI skills jumped to 34.2 percent in 2025, up from 11 percent a year earlier.

The 2026 Global AI Jobs Barometer is the most useful large-scale evidence we have seen this year on what AI is actually doing to jobs, because it measures listings and wages rather than sentiment. PwC's conclusion is not that AI is deleting work. It is that AI is sorting work into two tracks, and the tracks are diverging fast.

Professionalised versus democratised

PwC distinguishes roles that AI professionalises, where automating the routine parts concentrates the job around judgement and expertise (radiologists, recruiters), from roles that AI democratises, where the technology makes the work easier for non-experts to perform (IT service managers, medical secretaries). The first track is pulling away on every measure.

2026 Barometer finding	Figure
Job growth, professionalised vs democratised roles	~2x faster
Salary growth gap between the tracks since 2021	42 percent faster
Wage premium for AI-skilled workers, 2025	34.2 percent (from 11 percent)
Productivity growth, most vs least AI-exposed firms	40 percent higher
Headcount growth since 2018, most AI-exposed firms	52 percent vs 36 percent

Two findings deserve particular attention. First, the firm-level numbers undercut the assumption that AI-heavy companies shed people: the most AI-exposed firms grew headcount faster, 52 percent against 36 percent since 2018, while also posting higher productivity growth. Second, PwC's analysis of 2.4 million US entry-level listings shows AI-assisted junior roles are now seven times more likely to demand human-intensive skills such as leadership and face-to-face interaction. The bottom rung has not vanished; its job description has been rewritten.

What the divide means for employers

The scarce asset in the two-track market is not AI skill in isolation. It is people who can pair AI leverage with the judgement to know when the output is wrong.

For hiring, this cuts both ways. Recruiting for democratised roles gets easier and cheaper, but retention and differentiation get harder, because the same accessibility that widens your candidate pool widens your competitors'. Recruiting for professionalised roles is where the wage pressure lands, and the 34.2 percent premium says the market has already repriced them. Waiting for that premium to fall back has, so far, been a losing bet.

For development, the implication is sharper: the fastest way to move people onto the good track is to train judgement on top of AI fluency, not AI fluency alone. That matches what we see in workshops, where the durable gains come from teaching people to interrogate machine output in their own domain.

A talent agenda for the two-track market

• Classify your key roles: which are being professionalised, which democratised, and where your pay and progression models assume a world that no longer exists.
• Rebuild entry-level jobs deliberately around the human-intensive skills the data says they now require, rather than leaving juniors with the residue AI left behind.
• Budget for the AI wage premium in critical roles, and offset it by growing your own: internal upskilling is now cheaper than hiring the premium in.

The labour market is not waiting for anyone's five-year workforce plan. The divide is already in the listings data.

Related reading

• PwC press release: AI reshapes the global labour market into two distinct paths
• Hiring in the age of AI: the entry-level squeeze

---

Source: 2026 Global AI Jobs Barometer, PwC.