The labor economics of cognitive work turn on one variable: the marginal cost of executing a repeatable task. When that cost falls, employers restructure headcount before workers can reprice their skills. Goldman Sachs has now found that artificial intelligence is moving that margin in observable ways, with the bank's research pointing to employment pressure already visible across developed economies.
The significance sits in the tense. For years, the dominant frame around AI and labor has been prospective, treating displacement as a risk to be modeled and not yet measured. Goldman's researchers are now describing something present in the data. The signal is here; the forecast is behind.
Goldman's scope is developed economies specifically. That geographic focus is not incidental. Developed labor markets carry the highest concentration of knowledge-intensive roles, the functions that current AI tooling targets with the most precision. Employment tied predominantly to physical labor operates on a different substitution curve and faces the pressure later.
The question the finding opens is whether the pressure Goldman has identified represents the leading edge of a structural shift or a transitory adjustment. The bank has not, on the basis of this research, answered that. What it has done is move the conversation from conditional to observed.