There is a story being told about young people and artificial intelligence that goes something like this: the robots arrived, the entry-level jobs vanished, and a generation was left stranded before it even got started. It is a vivid story. It is also, in its present form, largely a fiction.
Youth unemployment has risen in several economies over the past two years. Graduate underemployment — the fate of the degree-holder serving coffee, not because they chose to but because nothing else materialised — is a genuine and worsening problem. These are real grievances, held by real people. But the leap from “young people are struggling” to “AI is the cause” skips over the evidence with remarkable confidence.
The data, such as it is, does not support the indictment. AI adoption in most industries is still concentrated at the experimental or assistive stage. The sectors shedding junior roles most aggressively — finance, law, consulting — were doing so before large language models became a boardroom talking point. The drivers then, as now, were hiring freezes, offshoring decisions, and post-pandemic overcorrections in headcount. To lay those decisions at the feet of ChatGPT is to let older, dirtier forces off the hook.
Consider what the statistical picture actually shows. When researchers control for the macroeconomic cycle — for interest rates that rose sharply and cooled growth, for a venture capital pullback that devastated the start-up ecosystem where so many graduates once found their footing — the AI “signal” largely disappears. The youth labour market is not uniquely wretched in places with high AI adoption. It is uniquely wretched in places with high housing costs, stagnant wage growth, and credential inflation: three structural failures that predate the current technology wave by decades.
None of this is to say that AI will not, eventually, reshape entry-level work in significant ways. It probably will. But “eventually” is doing an enormous amount of work in that sentence. History offers some guidance here. The spreadsheet was supposed to end the accounting clerk. The internet was supposed to hollow out junior roles in media, advertising, and publishing — and it did, but the timeline was measured in fifteen years, not eighteen months, and entirely new categories of work emerged alongside the destruction. Technological displacement is real. Technological displacement operating at the speed the current discourse assumes is not.
What is troubling is not the prediction itself but the function it serves. When young people’s difficulties are attributed to AI, a narrative emerges that is politically convenient for almost everyone. Employers avoid scrutiny for hiring decisions that reflect cost-cutting rather than capability. Policymakers sidestep questions about housing, student debt, and the adequacy of social infrastructure. And the technology industry, oddly, benefits too — there is no better advertisement for a tool’s power than a generation convinced it has already lost to it.
Young people deserve a clearer account of what is actually happening to them. The graduate who cannot find work commensurate with her qualifications is not, in most cases, being replaced by a model trained on trillions of tokens. She is more likely the victim of a housing market that has made mobility impossible, a credential system that overproduced degrees in the name of aspiration and underdelivered on the promise, and a corporate culture that has come to treat junior hiring as a liability to be minimised.
These problems demand political solutions: investment, reform, and the kind of patient institutional repair that does not generate a compelling headline. AI, by contrast, offers a clean villain — novel, slightly ominous, legible to anyone who has spent an afternoon arguing with a chatbot. The temptation to cast it as the antagonist in the story of generational decline is understandable. It is also a distraction from the harder work of understanding what has actually gone wrong.
The robots may yet take more than their share. But right now, they are being blamed for a theft that older, more familiar forces committed long before they arrived.
