There is a number doing a lot of work in career conversations this year, and it deserves to be handled carefully. Workers who can demonstrate AI skills command a wage premium of 62%, up from 57% a year earlier (PwC 2026 Global AI Jobs Barometer, Jun 2026). It is a striking figure, drawn from PwC's read of more than a billion job advertisements across 27 countries. It is also widely misunderstood, in a way that pushes people toward exactly the wrong preparation.
The misunderstanding is this: most people hear "AI skills" and picture code. They assume the premium is a reward for learning to build models, or for mastering a particular tool, and they conclude the door is closed unless they can do something technical. When you actually read what is inside the postings paying that premium, a different picture emerges.
What is inside the postings
PwC's own summary is blunt about it. Beyond coding, the report finds marketing, human resources, finance, and operations all hiring for AI fluency, often with no coding requirement at all (PwC 2026 Global AI Jobs Barometer, Jun 2026). The premium is not concentrated in engineering. It is spread across knowledge work, and it attaches to a capability more than a credential: the ability to use AI competently inside your actual domain.
More striking is what is being added to those jobs. The new tasks appearing in AI-exposed roles are 2.5 times more likely to lean on human skills such as empathy, judgment, and creativity than the tasks they replace (PwC 2026 Global AI Jobs Barometer, Jun 2026). As routine work gets absorbed, what is left over, and what gets paid for, is the part that requires a person to weigh, decide, and communicate. PwC's global AI officer frames the resulting labor market as a split into two tracks, one where AI amplifies human judgment and one where it does not (PwC, EME Outlook, Jun 2026).
So the premium is real, but it is not what the word "skills" implies. It rewards a hybrid: someone who knows their field well enough to direct an AI tool inside it, and who supplies the judgment the tool cannot. That is a very different thing to prepare for than a coding bootcamp.
Why the number is so large
A 62% average premium is large enough to warrant suspicion, so it helps to see the spread. The premium runs as high as 118% in some consumer-facing sectors and as low as 16% in government and public-sector work (PwC 2026 Global AI Jobs Barometer, Jun 2026). An average built from a range that wide is not a promise that learning a tool will raise your salary by two-thirds. It is a description of how differently the market is valuing AI fluency across industries, with the steepest rewards where AI most directly touches revenue.
The growth underneath the premium is genuine. Jobs requiring specific AI skills have grown roughly eight times as fast as the overall job market (PwC 2026 Global AI Jobs Barometer, Jun 2026). And the premium has climbed year over year, from 25% two years ago to 56% and now to 62% (PwC 2025 Global AI Jobs Barometer, Jun 2025; PwC 2026 Global AI Jobs Barometer, Jun 2026). The direction is consistent even if any single year's average flattens the reality.
What this means for your preparation
Read correctly, the 62% figure is more encouraging than alarming, especially for people who do not code. The market is not asking most knowledge workers to become engineers. It is asking them to become fluent, meaning able to hand routine work to a tool, check the tool's output with domain judgment, and keep the parts of the job that require a human to be present.
That reframes the preparation problem into something manageable. You do not have to abandon your field to capture the premium. You have to bring AI into the field you already know. A finance analyst who can direct a model through a reconciliation and catch where it went wrong is more valuable, not less, than one who cannot, and neither had to learn to program to get there.
There is a caution buried in the good news. Because the new work is 2.5 times more weighted toward judgment and empathy, the workers most exposed to displacement are those doing the routine portion without the judgment layer on top. The premium and the risk are two sides of one coin. The move that captures the premium is the same move that reduces the risk: stay close to the part of your work that requires a person, and let the tool take the part that does not.
The headline says AI skills pay 62% more. The postings say something quieter and more useful. What pays is judgment, applied inside a field you know, with a tool you have learned to steer. That is not a door closed to you. For most people reading this, it is a door that is easier to walk through than the number first suggests.