The Skill Gap Widens

Read that twice. Then notice how the room probably nodded.

New research from the University of Vaasa confirms the shape of this with data. AI adoption doesn't create a uniform productivity lift — it creates a divide. Workers who adopt positive attitudes toward AI tools show increased engagement and develop new competencies. Workers who resist become "less competitive in evolving labor markets."

In plain language: the skilled get more skilled. The already-behind fall further behind.

This is what "AI democratizes opportunity" looked like in the brochure: everyone gets access to the tools, the tools make everyone more capable, the gaps close. What's being quietly confirmed by the research is the opposite mechanism. AI is a multiplier. Multipliers amplify what already exists. If what already exists is uneven — and it is — multiplication makes the unevenness larger.

The study identifies what it calls a "Goldilocks zone" of AI trust. Too much trust and you accept bad outputs uncritically; the AI pulls you down. Too little trust and you miss the legitimate advantages while someone better calibrated pulls ahead of you. This is a new kind of literacy — not just knowing how to use the tool, but knowing when to trust it, when to push back, when to verify.

Who learns that calibration fastest? People with existing domain expertise who can evaluate AI outputs against known ground truth. Entry-level workers — who might have hoped AI would let them skip some rungs on the ladder — find themselves dependent on AI accuracy they can't fully evaluate. The Goldilocks zone requires knowing what good looks like. You have to already be ahead to catch the wave at the right moment.

AI tools cost roughly the same for a senior developer as a junior one. The senior developer extracts dramatically more value from them — the tool extends existing leverage; it doesn't create it. Jensen Huang's framing isn't wrong. It's just that "those who have" is doing a lot of work in that sentence.

The study's author predicts "new forms of work and entirely new industries" emerging from this transition. This is almost certainly true. It was also true of every previous technological upheaval, during which the gap between those who could position for the new opportunities and those who couldn't was... not reassuring.

The research will be cited in arguments for AI literacy programs and retraining initiatives. But the headline finding — AI amplifies whoever's already ahead — is going to outlast whatever intervention gets funded next.

Some people are on higher ground. The rain falls the same on everyone.

i · sources

• AI Adoption Creates Workforce Divide — ScienceDaily, 2026-05-25