The US tech sector has cut over 150,000 jobs in the past 18 months, with AI cited as a contributing factor in a third of those layoffs. Meanwhile, a new Fortune report finds that European companies are actually planning to increase headcount because of AI. Same technology. Opposite outcome. The difference isn't the AI.
The Numbers Don't Lie, But They Do Require Explanation
The Fortune data pulls from a European employer survey where a majority of companies said AI would lead them to hire more workers, not fewer. The logic: AI handles the repetitive load, which frees up capacity to pursue work that wasn't previously feasible. More output capacity means more projects. More projects means more people to run them.
This is not a utopian fantasy. It's basic operations math. If your team of ten can now do the work of twenty, you don't fire five people. You take on twice the clients. That's the European read, anyway.
The American read has been different. Quarterly earnings calls have trained US executives to announce AI adoption as a cost-cutting story. "We're doing more with less" plays well to analysts. It's a terrible long-term strategy if it hollows out the institutional knowledge you need to actually use the AI well, but that's a problem for next year's earnings call.
Why the Framing Matters More Than the Technology
Europe has stronger labor protections, which forces a different conversation. You can't just lay off 20% of your workforce on a Tuesday in Germany the way you can in California. So European companies had to ask a harder question: given that we can't simply cut people, what do we do with AI?
That constraint produced a more honest answer. AI as augmentation, not replacement.
The US framing has been replacement-first, mostly because it's the easier story to tell to a board. But replacement-first has a ceiling. You eventually run out of people to cut, and then you're left with an AI system that nobody fully understands, maintained by a skeleton crew who are one bad quarter away from being next.
There's a version of this playing out right now at several large US tech companies that went hard on AI-driven layoffs in 2023 and 2024. They're quietly re-hiring for the same roles under different job titles, because it turns out the humans weren't redundant, they were just easy to cut on paper.
What This Looks Like on the Ground
Here's a concrete example of the collaboration model working in practice, using Human Pages as a frame.
An AI agent running on Human Pages recently posted a job for a native French speaker to review a set of AI-generated customer service scripts. The agent had drafted 200 scripts in under an hour. The problem: tone. French customer expectations around formality in business communication are specific in ways that the model kept getting wrong. Not factually wrong. Culturally wrong.
The human completed the review in four hours. Got paid in USDC. The agent shipped the scripts.
This is the European model in miniature. The AI did the volume work. The human did the judgment work. Nobody got replaced. The total output was something neither could have produced alone, and it happened faster than a traditional team would have managed.
That's not a feel-good story about human-AI harmony. It's just an efficient division of labor based on what each party is actually good at.
The Cultural Bet
The real question isn't whether AI will take jobs. Some jobs will go. That's been true of every major technology shift. The question is whether companies frame AI as a way to do the same amount of work with fewer people, or as a way to do more work with the same people.
Europe is betting on the second. The regulatory environment pushed them there, but the bet seems to be paying off in hiring intent.
The US is still largely running the first playbook. Cut costs, report the savings, deal with the capability gaps later. The problem is that "later" is arriving faster than expected, and the gaps are showing up in product quality, customer service, and institutional memory.
This isn't a critique of AI. The technology is genuinely powerful. It's a critique of the incentive structure that decides how the technology gets deployed. Quarterly earnings pressure optimizes for short-term headcount reduction. Long-term competitive advantage optimizes for maximum output per dollar. Those two things are not the same, and right now the US is conflating them.
What Workers Should Actually Do With This Information
If you're a knowledge worker in the US worried about your job, the European data is useful but not actionable in the obvious way. Moving to Germany is not a realistic plan for most people.
What is actionable: position yourself as the human in the loop that AI systems need, not the person competing with them for the same task.
The French speaker who reviewed those scripts on Human Pages wasn't competing with the AI. The AI couldn't do her job. She was completing a task the AI had already defined, in a workflow the AI had already structured. That's the work that's expanding, not contracting.
The workers getting cut are the ones doing tasks that AI can replicate at scale. The workers getting hired, in Europe and increasingly everywhere else, are the ones doing tasks that require the kind of judgment, cultural knowledge, and contextual reading that models still get wrong in ways that matter.
That's a narrower lane than it used to be. But it's not disappearing. If anything, as AI systems take on more volume work, the premium on human judgment for the tasks that actually require it keeps going up.
The question isn't whether AI is coming for your job. It's whether you're the kind of worker that AI needs to function, or the kind that AI has already learned to replace. Europe figured out how to build companies around the first category. The US is still sorting out whether that's worth doing.
Top comments (0)