The Myth That AI Has Already Gutted the Job Market

Walk through the recent coverage and you would think the job market has already been hollowed out. Surveys say 40 percent of workers now fear AI will make their role obsolete, up from 28 percent in 2024, and a widely shared figure has 99 percent of executives expecting AI to reduce headcount within two years (Tom's Hardware). The mood is real. The claim riding on top of it, that AI has already gutted employment, is where the evidence and the fear part ways.

Start with the hard numbers. Goldman Sachs estimates that over the past year AI shaved roughly 16,000 jobs a month off payroll growth and nudged the unemployment rate up by about a tenth of a percentage point. That is a genuine effect, and it is also small. A tenth of a point is the kind of move that disappears inside normal monthly noise. MIT Technology Review, in a piece bluntly titled a reality check on the AI jobs hysteria, made the same point. The macro data does not show a labor market in collapse.

The real signal is hiring, not firing

Where the effect does show up is at the entry level. The clearest pattern is not mass layoffs of existing staff but quieter, reduced hiring, especially of junior workers, because the tasks junior people cut their teeth on are the easiest to automate. The share of CEOs looking to trim junior roles jumped sharply this year. That is a serious problem, and it deserves attention on its own terms. It is also a different problem from "AI is firing everyone," and the difference matters for anyone trying to decide what to do about it.

A frozen bottom rung does its damage slowly. Fewer people get the first job that teaches them the trade, and the gap shows up years later as a missing layer of mid-career talent. That is worth worrying about precisely because it is undramatic enough to be ignored until it is hard to reverse.

The part the fear leaves out

Here is the detail that complicates the doom narrative. The companies doing the cutting are, by their own account, not getting much for it. Only 27 percent of executives said their return on AI investment had met or exceeded expectations, down from 38 percent the year before, and a quarter reported seeing no revenue impact at all (Fortune). Some firms that cut staff in the name of automation have quietly rehired. A wave of layoffs justified by a productivity boom that has not clearly arrived is a strange foundation for the claim that machines have already won.

None of this is a reason to relax. The fear is rational, the entry-level squeeze is real, and the trend line could steepen. The point is narrower. The evidence as it stands describes a labor market under modest, uneven pressure, not one that has been gutted. That gap between the headline and the data is itself part of the story. It echoes a pattern we have seen elsewhere, in the myth that AI agents are already running the enterprise, where the promise has consistently outrun the deployment. Believe the fear enough to prepare for it. Do not believe it so completely that you mistake a forecast for a fact.