AI Displacement Is No Longer Subtext: The Attribution Shift Is the Story
TexTak has held a 73% probability on the first major AI-attributed layoff wave — and today's data makes a strong case that the threshold has already been crossed. A new analysis of Q1 2026 tech layoffs found that 47.9% of 78,557 cuts were explicitly tied to AI automation, up from under 8% of announcements in 2025. Jack Dorsey didn't bury the explanation in footnotes — he put it in a public memo. That behavioral shift in corporate communication is precisely what our forecast was measuring.
Our 73% reflected a specific thesis: that companies would eventually face enough investor pressure for demonstrated AI ROI that the reputational cost of public attribution would fall below the financial benefit of signaling AI-driven efficiency. We weighted the Block and Oracle data points heavily because they're not rumors or inference — they're named executives, named companies, named headcount figures, with explicit AI causation language attached. That's about as direct as evidence gets in this domain.
The Anthropic labor research adds a structural layer that's genuinely important. The 16% employment drop among workers aged 22-25 in AI-exposed fields, combined with the 14% decline in job-finding rates, tells us this isn't just a cyclical headcount reset — it's pipeline disruption. The mechanism matters here: it's not mass layoffs visible in the headline number, it's hiring freezes and role elimination at entry level that don't generate press releases but show up in cohort employment data. That's consistent with our model, which flagged attrition-based displacement as the dominant form even as explicit attribution increased.
The counterargument we take seriously is survivorship bias in the attribution data. The 47.9% figure comes from announced layoffs where AI was cited — but we don't have a denominator that includes quiet workforce reductions where companies said nothing about automation. It's possible that the most aggressive AI substitution is happening silently in sectors where public attribution still carries reputational risk (healthcare, government contracting, professional services). If so, the 47.9% figure understates the phenomenon but overstates the attribution shift in those sectors. Our forecast is specifically about public attribution, not displacement itself — so we're measuring the right thing, but we should be careful not to treat the tech sector's communication behavior as representative of the broader economy.
What would move us below 60%: evidence that the Q1 attribution spike was a one-quarter anomaly driven by a handful of high-profile executives (Dorsey, Oracle's leadership) rather than a durable communication norm. If Q2 layoff announcements revert to pre-2026 attribution rates despite continued headcount reduction, we'd need to revisit whether the threshold has genuinely been crossed or whether we're pattern-matching on outliers. We're watching Q2 earnings calls specifically — three or more Fortune 100 CEOs attributing net headcount decline to AI in prepared remarks would move us above 80%.