Companies are restructuring around a future that hasn’t fully arrived.
That’s the finding worth sitting with from a new NBER working paper, released around March 24, 2026, drawing on a survey of 750 CFOs conducted with Duke University and the Federal Reserve Banks. Commentary from workforce analyst Jason Averbook, whose Substack post confirmed the paper’s existence and described its findings, characterizes the projected cuts as often preemptive. Meaning: companies are making workforce decisions based on what they expect AI to do, not solely what it demonstrably does today.
The headline figures, reported via commentary on the paper, are striking. According to reporting on the NBER working paper, AI-driven job cuts in 2026 are projected at approximately 502,000 – roughly nine times the approximately 55,000 recorded in 2025. These are projections from a working paper, not a peer-reviewed final publication, and the specific figures were not confirmed through direct access to the primary document. Treat them as research projections requiring independent verification, not established facts. The NBER’s methodology, a survey of 750 CFOs through credible institutional partners, is a serious research basis. The conclusions warrant attention while the paper undergoes additional scrutiny.
A separate figure often cited alongside workforce projections: Goldman Sachs Research has estimated that approximately 300 million jobs globally are exposed to AI automation. That’s an exposure estimate, not a displacement projection, the distinction matters. Exposure means AI could affect those roles; displacement means it has or will. Most workforce AI research conflates the two, which inflates the apparent urgency. The NBER paper, as described, is more specific: it examines what CFOs say they plan to do.
The entry-level impact is where the data, if confirmed, is most consequential. The paper reportedly finds a 16% employment decline among entry-level workers in AI-exposed roles, according to reporting on its findings. That figure wasn’t confirmed from the page content retrieved in verification. It deserves direct investigation, entry-level erosion has compounding effects on workforce pipelines and skills development that senior-level cuts don’t produce.
What to watch: Direct access to the NBER working paper would allow verification of the specific figures and a clearer read on methodology. Watch for: coverage of the paper in T1/T2 economics or labor outlets; whether the Goldman Sachs 300 million figure gets updated or refined in upcoming research; and whether any Q1 2026 employment data from the Bureau of Labor Statistics reflects the pattern the CFO survey predicts.
TJS Synthesis: The preemptive framing is the editorial point that makes this data more than a headline number. Organizations aren’t waiting for AI to replace workers, they’re restructuring in anticipation of it. That behavioral pattern has implications that extend well beyond the job cut projections themselves. Hiring freezes, entry-level position eliminations, and reclassified roles don’t always appear in layoff trackers but cumulatively reshape workforce composition. HR leaders and workforce planners need to account for AI’s labor market effects before they’re visible in official statistics, not after. This brief connects to our full impact assessment deep-dive linked below.