AI pushes down pay for work you can standardize, such as translation, basic content or routine data tasks. At the same time, the premium for scarce technical expertise is rising: senior developers, DevOps engineers and SREs who build, secure and scale AI systems. IT salaries are splitting into two movements, not one.

You've been working in IT for a while. You see the headlines about AI taking over jobs. But your payslip tells a different story than the panic pieces.

The difference sits in what exactly you do. Those who can use AI to solve hard problems become more expensive. Those who do work AI can take over itself feel pressure on rates and assignments. That distinction determines whether you benefit this year or lose ground.

There is no national trend of "AI raises wages" or "AI lowers wages". The effect differs per sector and per task. That's confirmed by research from the SER (the Dutch Social and Economic Council), which states that in most cases AI has so far caused little direct wage change. Wage increases mainly occurred when work became more complex or people learned new skills.

"Not only work, but also income can be affected by the use of AI." - SER

Research by PwC (AI Jobs Barometer 2026) shows how large those sector differences are. In manufacturing and the TMT sector (technology, media, telecom), the wage premium for AI skills is highest. In financial services that premium is actually negative, around -8%. In government it's around -2%.

So: if you work at a scale-up with a modern stack, chances are higher that AI knowledge pushes your salary up. If you work in a sector with heavy regulation and slow adoption, you see less of that effect. If you want to understand salary development for IT professionals, the sector you work in matters just as much as your job title.

Freelancers in IT feel the effect of AI first and hardest. Research by the Boekman Foundation shows that almost 1 in 5 freelancers experience fewer assignments and lower income due to generative AI. Among translators that rises to about 1 in 3.

That number is not just about translators. The pattern applies to any type of work you can standardize well: simple scripting, basic data processing, generating content without domain knowledge. Almost 1 in 5 freelancers report an income drop over the past twelve months due to generative AI, often between 10 and 25 percent less.

Freelance IT rates in 2025 are not a given. They depend on how unique your work is. If you only do what a language model can also do, you'll see your assignments decline.

That's structural, not temporary.

It doesn't mean every freelancer should be afraid. It means you need to be able to explain why your work is not replaceable. Someone in a role as senior AI engineer builds the very systems that make other tasks obsolete. That's a completely different position than delivering executional work that AI copies.

The IT roles that get paid better thanks to AI knowledge are not necessarily the ones with "AI" in the title. It's about roles where you make AI systems reliable, secure and scalable. Think infrastructure, security and system design.

Research by SEO Amsterdam Economics shows that AI exposure between 2006 and 2018 was already associated with a slight increase in employment and labor income: plus 0.2% employment and plus 0.4% total labor income. Small, but consistently positive. And that was before the current AI wave.

SEO adds an important conclusion: "The positive effects of AI concentrate around specific skills, such as communication, organization and analytical ability." That is exactly where many senior tech profiles distinguish themselves from juniors.

Concretely, we see this in roles like an infrastructure software engineer who needs to run AI workloads at scale, or an information security analyst who has to make sure AI systems don't become a new attack surface. That's not replaceable work. That's work AI makes necessary.

Want to raise your salary or rate in a market that's changing because of AI? Then it's not about being able to use AI, but about being able to embed AI in a business-critical process. That's a different kind of skill than writing a prompt.

The Dutch employment agency UWV puts it sharply: "AI reduces demand for professions with many routine tasks and a strong focus on language or data." That is exactly the work freelancers are already feeling pressure on.

What does count: being able to design, secure and scale. Someone leading a team that builds AI functionality, like an engineering manager, combines technical knowledge with the responsibility to deliver that work reliably. Those combination roles are becoming scarcer, and therefore more expensive.

Also interesting: salary is not the only thing that counts. Research shows that a quarter of job seekers prefer a job with good AI facilities over a higher salary. That means employers who want to compete for hard-to-find professionals shouldn't only adjust the pay. The tools, the autonomy and the type of work matter too.

As a hiring manager or CTO, you're probably seeing two things happen at once. Some vacancies become easier to fill, because AI makes certain tasks obsolete. Other vacancies become harder, because demand for deep technical expertise is growing.

Rabobank research shows that Dutch workers themselves hardly expect job loss from AI: only 6% see major consequences for their own job and 7% for their own salary. That's a cautious picture, but it also says something: the labor market reacts more slowly than the technology itself.

Which means you can anticipate now, while competitors are still waiting. Whoever bases salary policy on old job profiles will miss the boat when attracting senior developers or DevOps specialists who combine AI skills with proven experience.

No theory. Hard execution.

That's exactly where a structured recruitment process makes the difference. Not gambling on whoever happens to apply, but searching with focus for the profile that fits where the market is heading.

IT salaries are not moving in one direction. AI pushes down pay for predictable, standardizable work and raises the value of scarce technical expertise. If you can demonstrate that your work is not easily automated, you stand stronger in any negotiation about salary or rate.

For employers, this means old job profiles and salary scales no longer automatically match what the market demands. Wondering whether your current approach still fits where tech talent is heading? Talk it through with someone who watches this market move every day.