The Attribution Wall Has Fallen: White-Collar AI Displacement Is No Longer Deniable
textak holds white-collar AI displacement at 73% — but that number requires an honest confession before we defend it. The original forecast target, 'first major layoff wave explicitly attributed to AI automation,' may already be resolved. Baker McKenzie, Meta, and the broader 100,000+ tech layoffs of 2026 have broken the attribution wall we predicted companies would maintain. This piece argues the forecast has resolved YES, explains why we held back longer than we should have, and defines what we're actually watching now.
Let's be precise about what has happened. Baker McKenzie announced layoffs of up to 1,000 employees in February 2026 and explicitly cited AI-driven efficiency as the driver — not restructuring, not macroeconomic conditions, not strategic repositioning. AI. Meta laid off 8,000 workers in May 2026 and publicly tied those cuts to AI infrastructure investment while reassigning 7,000 to AI-related roles. The SkillSyncer aggregation of 100,000+ 2026 tech layoffs identifies AI automation as the leading explicit factor across filings. These are not companies 'quietly' replacing roles. These are named, public, at-scale attribution events.
Our original resolution criteria — 'first major layoff wave explicitly attributed to AI automation' — did not specify a scale threshold, a named AI system, or a specific attribution language standard. That was our error. We had been treating events like IBM's earlier headcount commentary and Klarna's well-publicized AI displacement disclosures as 'indirect evidence' without ever defining why they fell short of the threshold. The honest reading is that we were moving the goalposts. The Baker McKenzie and Meta announcements, layered on top of sector-wide data showing 55,000 officially logged AI-related layoffs with modeled estimates of 200,000-300,000, satisfy any defensible definition of 'explicit attribution at scale.'
So why is the probability at 73% rather than ~95%+ resolved? Because we are transitioning this forecast. The 73% should now be read as our confidence that the original target has resolved YES — and we are flagging this publicly rather than quietly retiring the number. What we weight against full resolution confidence: the FinFlowMax modeling point that only 55,000 layoffs appear in official filings while the modeled real figure is 3-4x higher, which means the attribution behavior (companies publicly owning the cause) is still partially suppressed. The gap between 55,000 filed and 200,000 estimated is itself evidence that some portion of displacement is still being laundered as generic restructuring. The wall has cracked, not collapsed entirely.
The counterargument we take seriously: economists quoted in the SkillSyncer coverage explicitly debate whether AI is the primary cause or a convenient justification for restructuring that would have happened anyway. This matters. A company that planned to reduce headcount for margin reasons and named AI as the reason is not the same phenomenon as one that genuinely displaced roles that would otherwise exist. We cannot cleanly separate these with current data. That ambiguity is real, not a hedge. What would push us to formally close this forecast at resolved-YES: a single employer with 10,000+ headcount reduction that publishes role-level data showing eliminated functions being performed by named AI systems with measurable throughput parity. We haven't seen that yet. Baker McKenzie is close. We're watching for the next.