The Attribution Wall Is Real — And It's Exactly Why Our 70% Holds

Let's start with what the data does and doesn't show. The displacement numbers are real and significant — 95,878 workers across 249 named-company events by early May, a 92% increase in AI-role hiring paired with a 56% wage premium, and a college-educated cohort facing 2x the displacement risk of non-degree workers. That last data point is interesting and consistent with an AI-attribution thesis: knowledge work automation would theoretically hit bachelor's-degree holders disproportionately. But we should be precise here. Consistency is not confirmation. Prior tech downturns — 2001, 2008, 2022 — also disproportionately displaced knowledge workers, driven by offshoring, SaaS multiple compression, and over-hiring corrections. We can't claim the educational-skew pattern is independent confirmation of AI attribution without controlling for those alternative explanations. We're treating it as corroborating, not additive weight.

The more important question for our forecast is: why hasn't a Fortune 500 CEO simply said it? The answer isn't purely PR risk — it's legal architecture. Corporate counsel routinely advises against explicit AI attribution in layoff communications precisely because it creates WARN Act exposure, opens the door to disparate impact claims, and hands plaintiffs a causation argument in wrongful termination litigation. This isn't a solvable PR problem; it's a structurally suppressed disclosure environment. IBM's Arvind Krishna is the closest we've come — his 2023 statement that AI would replace roughly 7,800 back-office roles was explicit but forward-looking and tied to attrition, not a mass layoff event. That's a meaningful precedent, but it doesn't satisfy our resolution criteria: a 1,000+ position reduction with AI cited as the primary driver in official communications. IBM's statement was a hiring pause projection, not a displacement attribution. It doesn't resolve YES on our definition.

So what's driving the 70%? Three factors, and we'll be specific about their weights. First, investor pressure for AI ROI is intensifying — and at some point, the only credible way to demonstrate AI productivity gains is to connect them to headcount efficiency. That creates a disclosure incentive from the CFO chair that eventually overcomes the legal caution from the general counsel chair. We weight this moderately. Second, the sheer volume of 2026 displacement is making implicit attribution unavoidable — when a company eliminates 2,000 back-office roles the same quarter it announces a major AI infrastructure investment, the press reads it as AI attribution even without the word appearing in the 8-K. The question is whether a company will make that connection explicit in official language. We weight this as the primary driver. Third, regulatory pressure is building in a direction that could actually reduce the legal cost of attribution — if the EU AI Act's transparency requirements force disclosure of AI's role in workforce decisions (a live question under high-risk classification), the cost-benefit calculus for attribution shifts. We weight this as the smallest but most dynamic factor.

What would move us? Upward past 75%: a major company files an 8-K or issues a press release connecting a specific reduction event to AI productivity gains by name. Downward below 50%: legal analysis surfaces confirming that outside counsel at three or more major firms has issued standing guidance against AI attribution in WARN Act-covered layoffs, effectively institutionalizing the silence. The 70% doesn't yet fully account for the possibility that legal suppression of attribution is structurally durable rather than temporarily cautious — that's the gap in our model, and it's the part of this thesis that genuinely keeps us up at night.