Every AI transformation plan I read has the same silent assumption baked in. Somebody else picks up the humans it moves.
Usually the somebody else is the state. Retraining programmes, unemployment insurance, whatever safety net gets built when the displacement gets serious enough. I have started to believe that safety net is a story leaders are quietly relying on nobody to write.
The moment a plan named itself
A Chief People Officer at a Dutch insurer showed me her AI-augmented claims plan in July. Amsterdam office, forty pages, tight logic.
On page 22 there was a line labeled workforce transition, coordinated with national retraining channels. She said "coordinated" the way you say a thing that is already done.
I asked her what "coordinated" meant. She paused. Then she said: "I think it means we assumed."
That is the whole plan in one word. The biggest single cost of the transition, quietly offloaded onto a mechanism nobody in the room had confirmed exists.
She rewrote the line the next week. The number next to it did not get any smaller.
What the research says
Juliana Chueri and Petter Törnberg (Vrije Universiteit Amsterdam and University of Amsterdam, September 2026) read 1,514,950 parliamentary speeches from 33 national legislatures between 2023 and 2026. Two-stage classifier, human-validated at κ=0.947 for AI-work relevance, micro-F1 above 0.87 on the frame codes. They wanted to see what politicians actually say when they talk about AI and work.
Compensation accounts for just 2.3% of response-frame mentions, while enablement and investment dominate (55.2%), regulation and restriction follow (21.8%), and training (20.6%) appears at similar rates across families.
One in forty. That is how often, across three and a half years and 33 parliaments, elected politicians raise the question of compensating the people AI moves. The pattern holds across party families. The mainstream and radical right want unrestricted enablement, social democrats stay adoption-oriented, greens split evenly, and only the radical left pushes clearly for restriction.
The authors put the debate plainly.
The AI conflict thus concerns not compensation after disruption, but whether politics should enable technological change or govern its trajectory.
The argument in the room is speed versus governance. The question of who pays for the workers a transition moves is not on the agenda.
That leaves the moral work of the transition sitting inside the firm, whether the CFO wrote a number next to it or not. Budget for it now, or write the memo later explaining why you didn't.