Skip to content
Briefings are running a touch slower this week while we rebuild the foundations.See roadmap
AI: Jobs, Power & Money
15MAY

NBER: nine in ten firms untouched by AI

2 min read
15:55UTC

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.

← Back to GitLab signs the manifesto, Brussels backs outJump to analysis ↓
EconomicDeveloping
Key takeaway

Bosses expect AI to cut jobs while their own employees expect it to create them.

A survey of nearly 6,000 senior executives across the United States, United Kingdom, Germany, and Australia, published by the National Bureau of Economic Research, found that 90% of firms report no impact on employment or productivity from AI so far. 1 Sixty-nine per cent of the surveyed firms actively use the technology. Nine in ten see nothing happening.

The contradiction sits inside the forecasts. Executives at these firms predict a 0.7% employment decline over the next three years. Employees at the same companies predict a 0.5% increase. 2 One group expects cuts. The other expects growth. They work in the same buildings, use the same tools, and hold irreconcilable views of what comes next.

During the 1990s offshoring wave, management planned relocations years before workers learned their roles would move overseas. Approximately 3.4 million US manufacturing jobs were lost between 1995 and 2005. Workers could not prepare because they did not know. The NBER data, spanning four countries with different labour market systems, suggests this gap is structural, not cultural . If executives act on private bearish forecasts without informing staff, displacement will arrive as a shock rather than a managed transition.

Deep Analysis

In plain English

A research body surveyed nearly 6,000 bosses across the US, UK, Germany, and Australia and asked whether AI has yet affected hiring or productivity at their companies. Nine in ten said no. But the same bosses predict employment at their firms will fall slightly over the next three years. Workers at those same companies predict it will rise slightly. Someone is wrong. Given that bosses set hiring plans, their forecast is more likely to be self-fulfilling.

Deep Analysis
Root Causes

Information asymmetry within firms is the structural cause. Executives have access to strategic planning documents, vendor capability assessments, and board-level restructuring discussions that do not reach workers. The 1.2-percentage-point forecast gap (0.7% decline vs 0.5% increase) is more consistent with deliberate non-disclosure than with genuine disagreement.

The 69% active AI adoption rate combined with the 90% null employment impact suggests a deployment phase that is currently affecting task structure without reducing headcount. The NBER finding by Humlum and Vestergaard that LLM adoption produces occupational switching without net changes in hours or earnings supports this reading: impact is happening below the level of employment statistics.

Measurement lag is also structural. Employment surveys capture headcount but not task composition or hiring freeze effects. The Dallas Fed found displacement operating primarily through collapsed job-finding rates among workers under 25, a mechanism invisible to standard employment impact questions.

What could happen next?
  • Risk

    The executive-employee forecast gap may widen as deployment accelerates, producing a shock dynamic similar to 1990s offshoring where workers had no preparation time.

    Medium term · Medium
  • Consequence

    Policymakers relying on current employment statistics will underestimate displacement risk because the primary mechanism is hiring suppression, not firing, which appears later in official data.

    Short term · High
  • Meaning

    The 90% null result at 69% adoption rates confirms the technology is in a pre-deployment productivity phase; the employment shock, if it arrives, will be sudden rather than gradual.

    Long term · Medium
Sources:NBER (Yotzov, Barrero, Bloom, Bunn, Davis et al)
Mentions:United Kingdom →Prima →United States →NBER →Anthropic →Dario Amodei →
First Reported In

Update #3 · The AI jobs data contradicts itself

NBER (Yotzov, Barrero, Bloom, Bunn, Davis et al)· 28 Mar 2026
Read original →
Causes and effects
This Event
NBER: nine in ten firms untouched by AI
The largest cross-country executive survey reveals a dangerous information gap: employers expect job losses while their own workers expect gains.
Led to
Three in four never file claims
NBER finding that 90% of firms report zero AI impact persists because measurement systems cannot see the displaced workers.
Occurred 28 Mar 2026
Read story →
Different Perspectives
Entry-level and displaced workers globally
Entry-level and displaced workers globally
Challenger's 69% April hiring-plan collapse means the entry-level market contracted faster than announced layoff figures indicate. Workers aged 22-25 in AI-exposed occupations show a 16% employment decline since late 2022; the Stanford JOLTS analysis puts the real AI labour impact at 34 times the declared Challenger count.
Chinese courts and regulators
Chinese courts and regulators
The Hangzhou Intermediate People's Court upheld in April that employers cannot dismiss for AI cost reasons without offering retraining, confirming the Beijing court's December 2025 precedent under Labour Contract Law Article 40. Chinese workers now hold the only binding, judicially tested AI employment protections in any major jurisdiction.
Investors
Investors
Markets are rewarding the AI restructuring trade. Cloudflare reported record revenue alongside its 20% cut; the companies endorsing S.3339, a commission study bill with no enforcement mechanisms, are the same companies executing the restructurings the commission would study.
EU member states and Council
EU member states and Council
The Council's non-binding encouragement clause won the 7 May Digital Omnibus trilogue, dropping 18 months of work toward a binding employer AI literacy obligation. The outcome reflects the trade-off member states made: regulatory flexibility for employers over enforceable worker protections.
AI-era tech CEOs
AI-era tech CEOs
Cloudflare's Matthew Prince framed the 1,100-job cut as 'defining how a high-growth company operates in the agentic AI era', not a cost reduction. GitLab's Bill Staples published the most candid CEO-signed thesis of the cycle: agents will plan, code, review, deploy, and repair.
US tech workers and organised labour
US tech workers and organised labour
SAG-AFTRA's failure to win the Tilly tax, following WGA's settlement without AI training payment, confirms that organised creative workers cannot secure royalty mechanisms for AI-generated characters. For software workers, GitLab's 60-team structure eliminates the managerial co-ordination layer without replacing it with equivalent roles.