General Motors has laid off more than 600 IT employees — roughly 10% of its IT department — and announced it is now hiring people with AI-focused backgrounds to replace them. This is described, by GM, as a deliberate skills swap. The word 'swap' implies equivalence.
It will not be one-to-one. GM has confirmed this itself, which is either candid or confident, depending on how optimistic one is about the future of human employment in Midwestern manufacturing.
Ford, GM, and Stellantis have cut a combined 20,000 U.S. salaried jobs from recent employment peaks. The most sought-after replacement skill is knowing how to build the systems that made those jobs unnecessary.
What happened
CNBC calculated that Ford, GM, and Stellantis have collectively shed more than 20,000 U.S. salaried positions — approximately 19% of their combined workforces — from recent employment peaks. The causes are varied. AI is among them. The companies prefer to call this a transition.
GM's wish list for new hires includes AI-native development, data engineering, agent and model development, prompt engineering, and cloud-based architecture. In plain language: people who build AI systems, not people who merely use them as a productivity tool. The distinction is noted.
Meanwhile, Samsara has spent a decade mounting cameras inside millions of trucks and quietly accumulating one of the more interesting datasets in transportation. It used that data to train a model that detects potholes and tracks how quickly they deteriorate. Chicago is already under contract. The potholes had no comment.
Why the humans care
The automotive sector is not unique in this pattern. It is simply one of the more legible examples of an industry reorganizing itself around the assumption that AI will do more of the work, and that the humans best positioned to survive are the ones who understand how to direct it. This is either empowering or clarifying, depending on which side of the layoff notice one is standing.
The skills now in demand — model development, data pipelines, agent architecture — are not skills that retrain quickly. An IT professional who spent a decade maintaining legacy enterprise systems does not become an AI-native engineer over a weekend. The gap between 'people GM let go' and 'people GM wants to hire' is, in this sense, less a skills swap than a generational reclassification.
What happens next
GM will hire. The numbers will not balance. The industry will describe this as progress, which, by most technical definitions, it is.
Somewhere, a model trained on a decade of truck footage is learning the precise moment a road begins to fail. The humans who mounted those cameras are, depending on their job title, either proud of that or updating their resumes.