The Bank for Central Banks Says AI Is Freezing Hiring Before It Cuts Jobs

If AI is really coming for everyone's job, why isn't unemployment exploding? The Bank for International Settlements — the Basel institution that advises the world's central banks — just published its answer, and it matches something millions of job seekers already feel in their bones. Companies aren't firing people because of AI. For now, they've simply stopped hiring.

That diagnosis comes from BIS Bulletin No 130, "AI and the global economy: implications for central banks," published on July 28, 2026 by Iñaki Aldasoro, Leonardo Gambacorta, Enisse Kharroubi and Matthias Rottner. It is an eight-page briefing written for monetary policymakers, not workers. But inside its macro analysis sits the clearest official description yet of a labor market that looks calm in the statistics and feels frozen in real life.

A trillion-dollar bet, increasingly paid for with borrowed money

[Fact] The BIS documents that spending on data centres, semiconductors and related AI infrastructure has reached around 1% of GDP in the most exposed economies. [Estimate] Combined AI capital spending by the biggest hyperscalers runs on the order of $1 trillion across 2025 and 2026.

What concerns the BIS is not just the scale — it's the financing. [Fact] The bulletin describes this investment surge as increasingly debt-financed, leaning on public debt markets and private credit rather than the cash reserves that funded earlier tech buildouts.

Here is why that detail matters for your job. Debt creates deadlines. A company that borrowed billions to build AI infrastructure needs that infrastructure to produce visible returns, and the fastest visible return is lower labor costs. The way the AI boom is financed will shape how hard firms push automation — and how patient they can afford to be with human headcount.

The productivity payoff: potentially large, definitely late

[Fact] The bulletin's own summary is blunt: "The productivity payoff from AI, though potentially large, remains uncertain and uneven, across both sectors and countries."

[Claim] The authors point to a "productivity J-curve" pattern: when a general-purpose technology arrives, measured productivity often dips first — while firms reorganize, retrain and rebuild processes — before it climbs. Individual firms are already reporting gains from AI tools, but turning those micro wins into economy-wide numbers depends on how fast activity reallocates from firms that adopt badly to firms that adopt well.

[Fact] The United States shows the trade-off in miniature: sectors most exposed to AI are posting larger productivity gains, but partly at the cost of slower employment growth. Output per worker up, number of workers flat. If you work in software, finance or professional services, that sentence describes your last two years.

Two forces pulling your job in opposite directions

The BIS frames AI's labor market effect as a contest between two forces. Generative AI complements tasks that benefit from human input — judgment calls, exception handling, work that needs context and accountability. It substitutes for routine cognitive tasks that can be fully or partially automated.

[Fact] So far, the bulletin finds, actual displacement has been limited. But it flags early signs of substitution in specific tasks and occupations — and it names names: call centres and business centres.

That matches what our own tracking shows. Our occupation data puts customer service representatives at 76% AI exposure and 72% automation risk in 2026, up from 55% on both measures in 2023 — one of the steepest climbs among the 1,016 occupations we cover. See the full data

If you work in contact-center or back-office roles, the BIS just confirmed you are in the first wave — not the last one. That is frightening. It is also advance notice that most workers in history never got.

"Low hiring, low firing" — why the job hunt feels broken

Here is the part of the bulletin that explains 2026's strange job market. [Fact] The BIS reports that many firms remain in a wait-and-see phase on AI adoption, held back by regulation, implementation costs, organizational adaptation and doubts about AI reliability. [Claim] The authors note this caution is consistent with the "low hiring, low firing" dynamic observed across several countries.

Decode that for a worker and it means: if you have a job, you are relatively protected — firms aren't cutting yet. If you're looking for one — especially your first one — you are pushing on doors that quietly stopped opening. Unemployment stays low, so headlines stay calm. Meanwhile vacancy counts shrink, hiring processes stretch out and entry-level postings evaporate. The official statistics and your lived experience diverge, and both are telling the truth.

One warning from the BIS deserves emphasis: this freeze is a holding pattern, not a destination. [Claim] If AI tools keep expanding what they can do, the balance could shift from waiting toward substituting. The window in which firms hesitate is exactly the window workers should use.

Why central banks are nervous — and why that reaches your paycheck

[Fact] The bulletin's core message to policymakers: AI hits demand and supply at the same time. The investment surge pushes demand up today; the productivity gains push supply up tomorrow, on an unknown schedule. That combination blurs cyclical signals, making it harder for central banks to read whether the economy is overheating or softening — and harder to set interest rates correctly.

Rate mistakes are not abstract. Set rates too high because AI investment looks like overheating, and hiring freezes harden into layoffs. Set them too low because productivity gains mask price pressure, and inflation eats your raise. Either error lands on workers first. That is why the bank for central banks wrote this bulletin, and why it is worth reading even if you never touch a policy lever.

What you can actually do with this

The BIS wrote for central bankers. Here is the worker translation.

First, if your role is heavy on routine cognitive tasks — customer service scripts, standard document processing, first-line support — treat the bulletin's call-centre finding as your early notice. The complement side of the ledger is real: escalation handling, relationship repair, judgment on ambiguous cases. Move your week toward those tasks now, while your employer is still in wait-and-see mode.

Second, remember what wait-and-see means: most firms have not deployed yet. The gap between AI's demonstrated capability and its actual adoption is measured in years, and that gap is your preparation time.

Third, watch how your employer finances its AI push. Announcements funded by borrowing come with return deadlines; deadlines come with cost-cutting pressure. That is a better layoff predictor than any headline.

The honest caveats

This bulletin is an eight-page macro briefing, not an occupational study — it identifies directions, not job-by-job probabilities. And the hiring freeze has a competing explanation: years of elevated interest rates suppress hiring too, and the BIS itself concedes the signals cannot be cleanly separated. That is actually the bulletin's central point — AI has made the economy harder to read for everyone, central banks included. Anyone who tells you they know exactly how this plays out is selling something.

What the evidence does support: displacement is limited so far, concentrated in identifiable task types, and preceded by a visible hiring slowdown. Workers rarely get this much warning. Use it.

Sources

• BIS Bulletin No 130 — AI and the global economy: implications for central banks (Bank for International Settlements, July 28, 2026)
• Full bulletin PDF

---

This article was produced with AI-assisted analysis of the BIS bulletin and reviewed for accuracy. Figures for individual occupations come from our own tracking data, available on the linked occupation pages.