1 hour ago
6
Goldman Sachs Research put a number on what many workers already suspected: AI is eating jobs faster than it creates them. The bank’s April 2026 report estimates that artificial intelligence has reduced monthly US payroll growth by roughly 16,000 positions over the past year, enough to nudge the unemployment rate up by 0.1 percentage points.
That 16,000 figure is a net number, which makes the underlying churn even more dramatic. Goldman estimates that AI augmentation effects, think new roles created around AI tools, productivity-adjacent hiring, and complementary positions, have added about 9,000 jobs per month. The substitution side of the ledger, where AI simply replaces human labor, is considerably larger. Do the math and the gross displacement figure lands somewhere around 25,000 jobs per month before the offsetting gains.
Who’s getting hit hardest
The pain is not evenly distributed. Technology, management consulting, graphic design, and customer service are the industries feeling the most pressure, according to the Goldman research.
Gen Z and entry-level white-collar workers are bearing a disproportionate share of the displacement. This is particularly notable because prior waves of automation tended to hit blue-collar and manufacturing roles first. AI’s disruption pattern runs in the opposite direction, targeting the office park before the factory floor.
The numbers are already shifting
By June 2026, Goldman revised its net monthly job loss estimate downward to about 11,000 positions. That revision doesn’t mean AI slowed down. Instead, it reflects counterbalancing hiring in sectors like construction and other areas less exposed to AI substitution. The underlying drag from AI-heavy industries remained persistent.
Economist Joseph Briggs, who has been central to Goldman’s AI labor research, pegged the ongoing monthly drag at 10,000 to 15,000 jobs in sectors heavily exposed to AI advancements as of mid-2026.
To put these numbers in perspective, the US economy typically adds somewhere in the range of 150,000 to 250,000 jobs per month during expansions. A 10,000 to 16,000 job drag means AI is quietly trimming roughly 5% to 10% off headline payroll growth.
The augmentation side of the ledger
The 9,000 jobs per month on the augmentation side represent real positions: AI trainers, prompt engineers, machine learning operations staff, and roles where human judgment layers on top of AI output.
The problem is arithmetic. Nine thousand new jobs per month doesn’t offset 25,000 displaced ones. And the new roles often require different skills, sit in different geographies, and pay on different scales than the ones they’re replacing.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
English (US) ·