87,714 AI-Attributed Layoffs in Five Months: The Attribution Barrier Is Breaking
textak's forecast that a major AI-attributed layoff wave would hit 73% probability was built on one specific bet: that companies would eventually stop hiding AI displacement behind neutral language like 'restructuring' and start naming it. Today's data — 87,714 AI-attributed job cuts in the first five months of 2026, surpassing the combined total of the prior two years — is the closest thing to direct evidence we've seen that the attribution barrier is collapsing. Meta's explicit framing of 8,000 May layoffs as offsetting AI investment costs is the type of named attribution our forecast required. We're raising this forecast from 73% to 76%.
Let's be precise about what's happening here, because the distinction between the phenomenon and the attribution behavior has always been the crux of this forecast. We've argued for months that AI displacement was real but that companies were systematically avoiding public attribution — using AI as a quiet efficiency driver while framing headcount reductions as strategic realignment, geographic consolidation, or plain-old cost-cutting. The 87,714 figure matters not because displacement is now happening (it was already happening) but because employers are now citing AI as the reason in sufficient numbers that journalists and economists are tracking it as a named category.
The Meta case is instructive. Eight thousand workers cut, explicitly framed against AI investment costs — not 'rightsizing for the next phase of growth' but a direct statement that automation is substituting for headcount. This is the behavioral change our forecast was tracking. When a company with Meta's institutional weight makes that attribution publicly, it lowers the reputational cost for other firms to do the same. We're watching for a cascade effect in Q2 earnings calls.
Honestly, the counterargument that still keeps us up at night is the 'AI as cover' problem, which several economists named in the same reporting. Companies facing margin pressure for entirely conventional reasons — rising rates, slowing ad markets, post-pandemic overhiring corrections — may be opportunistically attributing cuts to AI because it sounds forward-looking rather than reactive. If 30% of these 'AI layoffs' are actually cyclical corrections wearing AI branding, then the phenomenon we're tracking (genuine automation-driven displacement) is being overstated by the attribution data. We weight this concern at moderate severity: the Meta case appears structurally genuine given the timing and specificity, but the aggregate figure almost certainly contains noise.
What would move us above 80%: a Fortune 100 non-tech firm — a bank, insurer, or retailer — explicitly attributing a layoff of 2,000+ workers to AI process automation in an earnings call with quantified productivity metrics. What would drop us below 60%: if Q2 earnings calls show companies reverting to neutral attribution language despite continued headcount reductions, suggesting the May spike was a temporary naming convention rather than a durable behavioral shift.