56% of 2026 Layoffs Cite AI Explicitly. The Attribution Threshold Has Been Crossed.
textak has held a 73% probability on the first major AI-attributed layoff wave, and today's data from Skillsyncer moves the needle in a direction that should unsettle skeptics: 56% of 2026 layoff announcements explicitly name AI as the driver, affecting 156,270 workers across 150 companies as of June 23. This isn't back-office whisper — Jamie Dimon named it from JPMorgan's podium in February. The forecast is not just directionally right; the specific behavioral threshold we were watching — public corporate attribution — is cracking open faster than we modeled.
Our 73% probability has always rested on a specific analytical bet: that the phenomenon (AI-driven displacement) and the behavior (companies publicly acknowledging it) would eventually converge, but that the behavioral lag would be the binding constraint. Companies had every incentive to describe headcount reductions in terms of 'restructuring' or 'efficiency' rather than hand critics a narrative about machines replacing workers. What changed?
The Skillsyncer data suggests the answer is investor pressure, not social conscience. When 56% of layoff announcements explicitly cite AI in the same earnings environment where AI ROI is the dominant shareholder question, the calculus flips: attribution becomes a feature, not a liability. Dimon's public statement is the clearest signal here — this is a CEO of the world's largest bank telling a congressional forum that displacement is already underway and moving faster than prior tech transitions. That's direct evidence of the behavioral threshold, not proximate evidence of conditions forming around it.
The counterargument that deserves respect: the Skillsyncer figure may be counting differently than our forecast intends. Our forecast targets a 'wave' — a concentrated, high-profile event where AI attribution is explicit and central, not cumulative tallies across 150 separate announcements. Oracle's 30,000-person reduction is one data point; whether it was framed publicly as AI-driven or operationally reframed matters to the resolution standard. We'd also note that the Jones Walker / Harvard Law finding — that AmLaw 100 firms are holding attorney headcount steady despite 100x productivity claims — is genuinely interesting, but it's a lagging indicator about strategic labor decisions in a professions context, not a proxy for whether the displacement wave has occurred in back-office functions. Those are different questions.
What would move us above 80%: a single company with 10,000+ employees issuing a public statement in an earnings call or 10-K that explicitly attributes a headcount reduction program to AI deployment, with a named number. Dimon's comments come close but describe reassignment rather than net reduction. What would drop us below 60%: Q3 earnings calls showing companies walking back attribution language under labor-relations pressure, or federal guidance discouraging explicit AI-displacement framing. We're watching both. For now, 73% feels like it may be lagging the data.