Three federal surveys, one 34-to-1 gap

Stanford Digital Economy Lab applied the JOLTS (Job Openings and Labor Turnover Survey) 3.1% hiring rate — the lowest reading since April 2020 — to the 158.6 million nonfarm workforce and calculated that AI is preventing approximately 950,000 to 1 million fewer annual hires compared with the 2023 pace. Against Challenger's 27,645 declared AI-attributed layoffs through March 2026, the real AI labour impact is approximately 34 times the declared count. Workers aged 22–25 in AI-exposed occupations show a 16% employment decline since late 2022; young software developers sit 20% below their 2022 peak.

The analysis reframes the displacement debate: announced layoffs are one-thirty-fourth of the actual labour impact. The dominant channel runs through hires that never happen, which no official US instrument currently tracks. Read full version ↓

The Bureau of Labor Statistics skipped its scheduled GenAI workplace publication on 14 April 2026. The Federal Reserve Bank of New York filled the vacuum the same day via Liberty Street Economics, publishing Survey of Consumer Expectations findings from November 2025 showing 39% of employed workers use AI tools, with access stratified 4.2x by income (66.3% earning over $200,000 versus 15.9% earning under $50,000) and 2.6x by education. Some 62% of workers expect AI to increase unemployment within 12 months. Only 15.9% of employers offer AI training; 11% actively prohibit AI tool use.

A federal measurement vacuum was filled by a regional reserve bank's household survey rather than by the Bureau of Labor Statistics itself, shifting the de facto US AI workforce measure from employer-reported to worker-reported data. Read full version ↓

A super PAC called Leading the Future, backed by OpenAI President Greg Brockman, Andreessen Horowitz and Palantir co-founder Joe Lonsdale, raised more than $125 million to defeat pro-regulation candidates in both parties in the 2026 US midterms. Total AI industry midterm spending is approaching $150 million. The fundraising surge followed within ten days of the Hawley-Warner letter becoming public. GZERO Media polling records 63% of Americans expecting AI to reduce employment and only 26% viewing AI positively; 50% of self-identified Republicans express concern about AI.

The feedback loop from AI displacement to regulation is being pre-empted at the electoral stage rather than the legislative one, with industry money targeting the measurement-minded legislators who could compel better federal data. Read full version ↓

The UK AI Security Institute (AISI) published an independent evaluation of Anthropic's Claude Mythos Preview on 15 April 2026. On isolated CTF (capture-the-flag) tasks Mythos scored above 85%, with GPT-5.4, Claude Opus 4.6 and Codex 5.3 falling within 5–10 percentage points — no single-task superiority. In AISI's 32-step 'The Last Ones' benchmark, Mythos autonomously completed an operation the Institute estimates would take a trained human roughly 20 hours, confirming a genuine attack-chaining capability distinct from the zero-day marketing claims.

The first external confirmation that the Treasury-Fed emergency convening on 8 April was warranted on capability grounds rather than on Anthropic's marketing. Attack chaining is the capability most directly relevant to autonomous task completion, and therefore to white-collar workforce displacement. Read full version ↓

The Islamic Revolutionary Guard Corps (IRGC) explicitly named Stargate UAE — the $500 billion OpenAI, SoftBank and Oracle joint venture — in a 1 April military targeting video. Amazon Web Services subsequently declared 'hard down status for multiple zones' after Iranian missile strikes on infrastructure in Bahrain and Dubai. Oracle's Dubai data centre had previously been struck. Iran named Nvidia and Apple by name in the same video. The Oracle data centre programme ($156 billion), funded in part by the 31 March termination of up to 30,000 workers, now includes physical assets in an active war zone.

The first recorded instance of a state military singling out AI data centre capex as a strategic target. The workers displaced to fund the infrastructure now also carry the physical war risk against it. Read full version ↓

Anthropic published a 244-page Alignment Risk Update for Claude Mythos Preview on 7 April 2026, abandoning AI Safety Level (ASL) capability thresholds in favour of autonomy-focused threat models. Project Glasswing was expanded to include Broadcom, CrowdStrike, Nvidia, Palo Alto Networks and Cisco alongside the original twelve partners. Over 99% of the vulnerabilities Mythos discovered remain unpatched, with coordinated disclosure still in progress.

A methodology shift that forces all Glasswing partners to rebuild their internal risk frameworks mid-deployment, while over 99% of Mythos-discovered vulnerabilities remain unpatched during coordinated disclosure. Read full version ↓

The Bank of England Financial Policy Committee (FPC), in its April 2026 record, asked the Bank of England and the Financial Conduct Authority for further work on agentic AI in payments and financial markets, noting that three-quarters of UK financial firms now deploy AI and that systemic risk from agentic deployment is 'likely to increase rapidly'. The Treasury Committee has called for AI-specific stress tests and clearer FCA guidance by end of 2026.

The UK has placed AI capability risk inside a formal financial stability framework. The US equivalent, Bessent and Powell's 8 April convening, remains ad hoc with no published follow-up. Read full version ↓

The Office for National Statistics recorded UK vacancies unchanged at 721,000 for a sixth consecutive publication, with payrolled employees down 96,000 year-on-year and real wage growth at 0.4% against 3.8% nominal. The ONS has no AI-specific breakdown of the labour market data.

Britain's labour market is frozen rather than recovering. The ONS has no AI-specific breakdown of its data, leaving the UK with the same measurement blindness as the US just as the Bank of England has raised systemic flags about AI in financial markets. Read full version ↓

The EU Digital Omnibus second trilogue is scheduled for 28 April 2026, twelve days from publication. The employer AI literacy obligation stripped by Parliament on 26 March remains contested entering negotiations. The Cypriot Presidency is aiming for political agreement by May, with Official Journal publication targeted for July and the high-risk employment deadline fixed at 2 December 2027.

The trilogue will determine whether EU workers gain a statutory right to AI transparency at work by December 2027. If the literacy clause survives, European employees will hold enforceable AI disclosure rights the US equivalent still lacks. Read full version ↓

Oracle's Massachusetts WARN Act filing remains absent as of 13 April 2026. The 60-day clock from the 31 March cuts expires around 30 May. Oracle's Burlington offices remain unrepresented in any state disclosure, and law firms are still investigating potential violations. Prior filings covered Washington state (491 positions) and Missouri (539) — covering fewer than 4% of the up to 30,000 affected workforce.

If Burlington produces no filing by 30 May, the precedent is set for a 96%-invisible workforce reduction at scale under the federal WARN Act template. Read full version ↓

JPMorgan Chase CEO Jamie Dimon confirmed at the bank's February 2026 investor meeting, as reported by CNBC, that the bank has 'displaced people from AI' and offers them other jobs. JPMorgan committed $600 million annually to retraining and tied engineer performance reviews to AI tool adoption across 65,000 staff. MetaIntro research circulating the same period found only 6% of companies globally are actually reskilling workers for AI. JPMorgan is one of the twelve original Glasswing partners with privileged Mythos access.

JPMorgan is the only major bank publicly acknowledging AI-driven internal displacement while simultaneously holding privileged Glasswing access to the restricted frontier model that triggered the federal emergency convening. Read full version ↓