A survey of 750 CFOs finds AI-driven layoffs will be nine times higher in 2026 than 2025, yet a parallel study of 6,000 executives shows 90% of firms report zero employment impact so far. The gap between what companies plan and what they measure defines a week in which the EU voted to delay workplace AI rules by 16 months, US senators split into competing camps, and the first AI disclosure law in the world produced no data at all.
Firms intend nine times more AI job cuts in 2026 while 90% report zero impact so far.
Seven hundred and fifty chief financial officers told the Atlanta Fed they expect AI-attributed layoffs to be nine times higher in 2026. The number is smaller than it sounds.
Atlanta Federal Reserve survey of 750 CFOs projected AI-attributed job cuts in 2026 to be nine times higher than 2025, totalling approximately 502,000 roles or 0.4% of the US workforce.
The first large-scale employer survey to quantify 2026 AI displacement projects 502,000 cuts, reframing the debate from panic to proportion.
A multinational survey of 6,000 executives found most companies see no employment effect from AI. Inside those same firms, bosses and workers hold opposite forecasts.
Nearly 6,000 executives across 4 countries found 90% report no AI employment impact even though 69% use it. The same executives predict a 0.7% employment decline over 3 years; their workers predict a 0.5% increase.
Bosses set hiring plans and access strategic documents workers do not. That 1.2-percentage-point forecast gap points more to deliberate non-disclosure than genuine disagreement.
The European Parliament voted 101 to 9 to push high-risk AI employment rules to December 2027. Buried in the amendments: employer AI literacy obligations have been removed entirely.
The European Parliament voted 101 to 9 on 26 March to delay the AI Act's high-risk employment rules from August 2026 to December 2027. A less-noticed provision removed employer obligations on staff AI literacy.
Europe's AI Act contained the only binding framework requiring employers to understand AI tools deployed against their workforce. Companies now face no such obligation until late 2027.
A 135-country study found women hold nearly twice the share of highest-risk AI-exposed jobs globally. In wealthy nations, the gap is almost threefold.
A 135-country study by the International Labour Organisation and World Bank found 4.7% of women in the highest AI exposure category, versus 2.4% of men. In high-income countries the gap widens to 9.6% versus 3.5%.
Clerical and administrative roles carry the highest AI substitution scores and are held more by women. 90% of projected UK AI job growth falls in professional roles instead.
New York required companies to disclose AI's role in mass layoffs. After a year, 162 companies covering 28,300 workers attributed zero cuts to AI.
United StatesNew York's updated WARN Act required AI disclosure in mass layoff notices. After nearly a year, 162 companies covering 28,300 workers filed (including Amazon and Goldman Sachs) and not 1 cited AI.
The penalty runs $500 per day, trivial for billion-dollar firms. Proposed legislation raising that to $10,000 per violation has not advanced. Without meaningful enforcement, AI disclosure remains optional in practice.
A bill to ban all new AI data centre construction until Congress passes worker protections. It will not pass. It was not designed to.
United StatesSenators Bernie Sanders and Alexandria Ocasio-Cortez introduced the AI Data Centre Moratorium Act on 25 March. The bill bans data centre construction pending worker protection laws; US household electricity costs rose nearly 7% last year.
The bill has no path in a Republican-controlled Congress. Its purpose is to force competing proposals to account for energy costs alongside labour displacement.
Nine senators across both parties wrote to federal agencies demanding expanded data collection on AI's workforce effects. It is the first evidence of a durable centre on AI labour policy.
9 US senators wrote to the Bureau of Labour Statistics and the Census Bureau in March, urging expanded AI workforce data collection. Senators Mark Warner and Josh Hawley led; 7 signatories joined across committees.
Federal agencies can act without new legislation. Current BLS occupational coding predates generative AI and cannot distinguish AI-augmented from AI-replaced roles.
Azure grew 40%. Revenue beat consensus by $2 billion. The first earnings test of the AI capex thesis returned a passing grade.
Microsoft posted fiscal Q1 2026 revenue of $77.7 billion, beating consensus of $75.6 billion. Azure grew 40% (the fastest rate in over a year) and monthly active AI feature users hit 900 million.
This is the 1st result after the Big Five committed $650-690 billion to AI infrastructure. Microsoft converts AI capex to cloud revenue more directly than peers. Meta reports April 29.
One CTO for AI product development. One for enterprise trust. The first tier-one software company to decompose its top technical role for the AI era.
Atlassian replaced departing CTO Rajeev Rajan with two AI-specialised chief technology officers — Taroon Mandhana (CTO of Teamwork, overseeing Rovo AI) and Vikram Rao (CTO of Enterprise and Chief Trust Officer) — as its 1,600-person layoff took effect.
Atlassian's restructuring signals that the CTO role itself is being reshaped by AI, not just the workforce beneath it.
A UK banking group announced layoffs one day after a short-seller flagged a billion-pound liability. AI was named as a reason. It may not be the real one.
United StatesUK banking group Close Brothers announced 600 job cuts on 23 March, citing AI and offshoring in an £85 million cost programme. The announcement came 1 day after a short-seller flagged a £1.2 billion liability.
Citing AI the day a major liability surfaces fits the AI-washing pattern. New York's $500-per-day non-disclosure penalty shows weak enforcement changes nothing.
More than 1,000 jobs gone. The CEO pointed to declining Fortnite revenue, not AI. A rare explicit denial in a quarter defined by AI attribution.
United StatesEpic Games cut more than 1,000 jobs on 24 March, roughly 20% of its workforce. CEO Tim Sweeney explicitly denied AI played any role, citing declining Fortnite engagement and operating losses.
In a quarter where 1 in 5 tech layoffs cite AI, that explicit denial is notable. Epic spent more than it earned. Not every tech cut is an AI story.
Fifteen of eighteen industry sectors posted vacancy declines. Real wage growth sits below 1%. Workers losing ground to inflation have less runway to absorb disruption.
UK job vacancies fell to 721,000 in the 3 months to February, down 9.5% year on year. The ONS reported on 18 March that unemployment reached 5.2% and real wage growth was just 0.4-0.5%.
Fewer vacancies and near-zero real wages leave workers with little financial runway. The UK labour market carries limited slack going into AI-driven restructuring.
The government's own figures show 90% of net new jobs will be professional-tier. The workers being displaced will not fill them.
The UK Government projected in January that AI-direct employment will grow from 158,000 jobs in 2024 to 3.9 million by 2035. 90% of that net growth falls in professional or associate professional roles.
Workers most exposed to displacement (clerical and administrative staff) will not fill those 3.9 million positions. No current policy addresses that mismatch.
Weekly claims dropped to 205,000 while insured unemployment hit a two-year low. The number contradicts the structural weakening visible in payroll data.
US initial jobless claims fell to 205,000 for the week ending 14 March, with insured unemployment at 1,819,000, the lowest since May 2024.
Other data cut against this: nonfarm payrolls fell 92,000 in February and tech layoffs passed 59,000 this year. Claims and payrolls tell opposite stories about the same economy.
Escalating. The CFO intent data (nine times 2025 levels) is set to convert into announced layoffs through Q2-Q3 2026, while every institutional mechanism that could moderate or measure displacement has moved in the wrong direction this week: EU rules delayed, disclosure law nullified by inadequate penalties, federal data collection not yet mandated. The only structural counter-signal is the 55% regret rate among leaders who already cut — but market incentives for AI-attributed restructuring remain unchanged.
