AI's Two-Track Job Market: Inside PwC's 2026 Global AI Jobs Barometer - July 1, 2026

Welcome to the DX Today Podcast, your daily deep dive into the AI ecosystem. I'm Chris. And joining me as always is Laura. Thanks, Chris.

And I'm genuinely excited about today's topic because it takes one of the most anxious questions of our era and finally answers it with actual data rather than pure speculation. You're talking about the jobs question, right? The one everybody at every dinner table keeps asking about whether artificial intelligence is quietly coming for their paycheck and their whole career. Exactly that question.

And the reason I love the new PWC Global AI Jobs Barometer is that it stops guessing and instead reads the labor market directly through more than 1 billion actual job advertisements. Hold on, 1 billion job ads is a staggering number. So before we get into the findings, help me understand the scale and the method behind a study that large and ambitious. So PWC pulled together more than 1 billion job postings across 27 countries and territories.

Then fuse that with company financial data and detailed occupational task data to see how AI is actually reshaping work. That combination is important because a lot of previous studies just counted job losses or made scary macro predictions. But this one is looking at wages hiring and the specific tasks inside each role. Right.

And the headline finding is that AI is not producing one uniform future for everyone. It is instead splitting the labor market into two very distinct tracks that PWC calls a new divide. A new divide, okay, that phrase is doing a lot of work. So walk me through what those two tracks actually are and why the distinction matters so much for ordinary workers listening right now.

The first track PWC calls professionalized roles, where AI absorbs the routine, repetitive parts of the job, which frees the human to lean harder into judgment, expertise, and the genuinely difficult decisions machines still cannot make. So in that professionalized track, the technology is essentially acting like a very capable assistant that clears the busywork off your desk so you can focus on the harder, higher value thinking. Precisely. And the second track is what they call democratized roles, where AI makes a job so much easier that people without deep specialized training can suddenly step in and do it competently.

On the surface, that democratized track actually sounds like good news because lowering the barrier to entry means more people can access work that used to require years of expensive training and credentials. It does sound good, and in some ways it is, but here is the uncomfortable twist that the data reveals. Because when a job becomes easy for almost anyone to do, its economic value tends to fall. Ah.

So this is the classic supply and demand story playing out. Where if the pool of people who can perform a role suddenly widens, the leverage and the pay for that role gets competed away. You've got it exactly. And the numbers make this painfully concrete because professionalized roles are seeing twice the growth in available jobs and 42% faster salary growth than the democratized ones.

Wait, let me make sure I heard that right, because 42% faster salary growth is not a rounding error. That is a genuinely enormous gap opening up between two groups of workers. It is enormous, and to make it tangible, PWC gives examples. So think of radiologists or recruiters on the professionalized side versus roles like IT service managers or medical secretaries on the democratized side.

That radiologist example is fascinating to me. Because for years, radiology was the poster child for the job that AI was supposedly going to completely automate away and eliminate entirely. And instead, the opposite happened because AI took over the routine image screening while the radiologists moved up into complex diagnosis, patient consultation, and the judgment calls that carry real medical and legal weight. So the lesson is not that AI came for the radiologist, but that AI came for the routine tasks inside radiology.

And the human quietly climbed to the more valuable part of the work. That reframing is the entire point of the report because it moves us away from the lazy headline that AI destroys jobs and toward the more accurate picture that AI redistributes the tasks within jobs. Now I want to push on this a little because a skeptic could say that PWC is a consulting firm with an obvious commercial interest in telling everyone that AI is wonderful for the economy. That is a fair challenge, and I take it seriously.

But the strength here is the sheer size of the data set. Because 1 billion real advertisements across 27 countries is very hard to spin into a convenient narrative. That is true. Real job postings with real salary bans are about as close to ground truth as you can get, since companies are putting actual money behind those listings rather than opinions.

And the company level numbers reinforce it, because firms most able to use AI saw headcount grow 52% versus 36%, and wages grow 24% versus 17%, compared with the least exposed firms. So the companies leaning hardest into AI are simultaneously hiring more people and paying them more, which really complicates the simple story that automation just means fewer humans doing the same work for less. It complicates it enormously. And there is one figure that genuinely made me stop and reread it because the so-called superstar companies most exposed to AI achieved labor productivity gains of 163%.

163% is almost hard to picture. So help our listeners understand what a productivity jump of that magnitude actually feels like inside a real functioning business day-to-day. Think of it this way: roughly the same team of people producing more than two and a half times the output they did before, which means the economic pie each worker helps bake is dramatically larger. And presumably that expanding pie is exactly why those companies can afford to hire more and pay more at the same time, rather than being forced to choose between growth and headcount.

Exactly. Productivity growth is what breaks the zero-sum thinking. Because when each worker generates far more value, the business has both the money and the incentive to expand its team rather than shrink it. Okay, so let's talk about the people listening who are wondering what any of this means for them personally, because the natural next question is how you land on the good side of this divide.

The clearest signal is AI skills, because jobs that require specific AI skills are growing about eight times faster than the overall market. 69% growth against just 9% for jobs as a whole. Eight times faster is the kind of gap that should make anyone sit up. Because that is no longer a gentle trend.

That is a wave that will reshape careers within just a few years. And there is a direct financial reward attached because the average wage premium for AI skills has now climbed to 62%, meaning workers who can genuinely wield these tools command dramatically higher pay. A 62% premium is remarkable, but let me play devil's advocate again. Because sometimes these skill premiums are temporary bubbles that collapse the moment the broader workforce catches up, and the skill becomes common.

That is a smart caution and it may compress over time. But the deeper findings suggest the premium is not just about tool knowledge, it is about the human skills that the tools make more valuable. That is an intriguing distinction. So unpack that for me, because I think a lot of people assume the future belongs purely to the coders and the prompt engineers who can talk to the machines.

Here's the surprising part because the new tasks being added to AI-exposed roles are two and a half times more likely to rely on skills like empathy, judgment, and creativity rather than raw technical ability. So the counterintuitive punchline is that as the machines get better at the technical and the routine, the distinctly human capabilities like reading a room and exercising wisdom become the scarce and valuable things. That is exactly it. And it flips the anxious narrative on its head.

Because the safest ground is not competing with AI on speed or memory, it is doubling down on what remains stubbornly human. There is one finding in here, though, that genuinely worries me, and it is the part about entry-level roles. Because young people just starting out are often the most vulnerable in any big economic shift. You are right to flag that, because the US data shows AI-exposed entry-level roles are now seven times more likely to require traditionally senior-level skills such as judgment and leadership than they were before.

That that is a real problem, because if the entry-level job suddenly demands senior level judgment, then how on earth is a new graduate supposed to get their first foot onto the professional ladder at all? It is the central tension of the whole report because AI is quietly eating the simple starter tasks that used to be how young workers learned the ropes and slowly built up their professional judgment. So we risk a situation where the bottom rungs of the career ladder are being sawn off, leaving employers demanding experience that the next generation has no obvious way to actually go and acquire.

And that is precisely where I think schools, universities, and employers have to step up. Because if AI removes the traditional training ground, then we need to deliberately build new on-ramps to replace what is disappearing. That feels like the honest takeaway because the technology itself is neutral. But the social choices we make around training, education, and hiring will decide whether this divide becomes a fair ladder or a locked gate.

Beautifully put, and it is why I keep coming back to the phrase a new divide, because divides are not destiny, they are the result of decisions, and decisions can absolutely be made differently. Before we wrap, I want to ask about the global picture, because this barometer spans 27 countries. And I wonder whether this two-track divide looks the same everywhere or shows up unevenly across regions. The broad pattern holds across markets, but the intensity varies with how deeply each economy has actually adopted AI.

So the countries and sectors furthest along the curve are where this new divide is opening fastest and widest. That makes me think the divide is not only between individual workers, but potentially between entire economies, where the nations that invest early in AI skills pull ahead while the slower adopters quietly fall behind. That is the sobering macro version of the story. And it is exactly why so many governments are now treating AI skills training as genuine economic infrastructure, on par with roads, broadband, and the electrical grid itself.

So if you had to compress this entire barometer into one piece of advice for someone listening on their commute right now, what is the single thing you would tell them to go and do? I would say learn to work with these tools directly and deliberately, but pair that with sharpening the human judgment, communication, and creativity that no model can replicate, because that combination is where the durable value lives. That is genuinely useful and grounded advice. And I appreciate that this whole conversation moved us past the tired fear headlines into something far more actionable and honestly a lot more hopeful about the future.

Me too. And the big reframe I want everyone to carry away is that I is not simply a job destroyer, it is a task redistributor. And where those tasks land is still very much up to us. That's all for today's episode of the DX Today Podcast.

Thanks for listening, and we'll see you next time.