US-first · AI Impact Monitor

AI Jobs & Work Monitor

A living view of how AI is changing work in the real economy — combining structural task exposure with observed AI use, hiring pressure, entry-level access, wages and skill demand. The goal is not to predict whether a job “dies.” It is to show what is changing, how strong the evidence is, and what to watch next.

Check an occupation Get weekly briefing

Occupation lookup · two independent exposure views

How much is AI touching the job — and how exposed is it relative to others?

Search observed professional usage from Anthropic alongside the official BLS five-source category. One measures current task coverage; the other compares relative exposure across occupations. Both are evidence — neither is a job-loss probability.

Explore occupation intelligence

Start with a real job, not a generic fear score.

Each page separates observed AI exposure from labor-market outcomes and shows exactly where the evidence comes from.

15-125228.8% observed

Software developers

BLS relative exposure: Very high · Neither measure is a job-loss probability.

Open intelligence →

13-201134.8% observed

Accountants and auditors

BLS relative exposure: Very high · Neither measure is a job-loss probability.

Open intelligence →

23-101116.7% observed

Lawyers

BLS relative exposure: Very high · Neither measure is a job-loss probability.

Open intelligence →

43-405170.1% observed

Customer service representatives

BLS relative exposure: Very high · Neither measure is a job-loss probability.

Open intelligence →

29-11416.0% observed

Registered nurses

BLS relative exposure: High · Neither measure is a job-loss probability.

Open intelligence →

U.S. labor-market baseline · stored

What is the labor market doing before we attribute anything to AI?

These official BLS series are context, not proof of AI causality. They give the monitor a clean economic baseline against which occupation-specific and AI-adoption signals can later be compared.

Jul 2026stored

4.1%

U.S. unemployment rate

-0.1 pp vs. prior month

Official source ↗

Jul 2026stored

158.9M

Total nonfarm payrolls

-0.02M vs. prior month

Official source ↗

Jul 2026stored

7.27M

Job openings

-0.09M vs. prior month

Official source ↗

Interpretation rule: a weaker labor market is not automatically an AI effect. AI-specific conclusions require a second layer — task exposure, observed AI use and occupation-level divergence.

Current AI-related evidence

U.S. office and administrative support employment is now projected to fall 4% by 2035.

AI-related pressure is now part of the official U.S. long-run employment baseline, not only a finding in vendor or academic exposure models. The projection still does not prove that AI alone will cause the decline; BLS incorporates several structural and economic forces.

What changed

The new BLS 2025–35 projections show office and administrative support occupations shedding 752,100 jobs — the largest loss of any major occupational group. BLS says continued workflow automation, including AI tools, is likely to reduce demand for several occupations in the group.

Who may feel it

Office and administrative support workers, employers, workforce planners and training providers — especially where work centers on repetitive information processing, documentation or customer handling.

Watch next

Occupation-level employment, hiring, wages and openings for customer service representatives, data entry keyers and administrative assistants; annual projection revisions; and the new BLS AI exposure categories.

Loading verified occupation evidence and stored U.S. labor context…

Source architecture

Built from open evidence, not a single black box.

The first U.S. version is designed around public or openly licensed sources so the core product can remain inexpensive, reproducible and highly automated.

US · open-data

O*NET

Occupation taxonomy, task statements, skills and task importance

CC BY 4.0Periodic releases

Source ↗

US · public-api

U.S. Bureau of Labor Statistics

Employment, wages, projections and official five-source AI exposure categories

U.S. government public dataMonthly / annual by dataset

Source ↗

US · public-api

U.S. Census Bureau — BTOS

Business adoption and operating-condition evidence, including AI use

U.S. government public dataHigh-frequency releases

Source ↗

Global/US · open-data

Anthropic Economic Index

Observed AI task usage, automation and augmentation evidence

CC BY data / MIT code where publishedVersioned dataset releases

Source ↗

US · public

AI Risk Index

Optional external structural occupation exposure, substitution and augmentation scores

CC BY 4.0 for scores and derived data; attribution + version requiredVersioned index releases; API availability may vary

Source ↗

US · public

Stanford Digital Economy Lab — Canaries Dashboard

Early-career employment evidence in AI-exposed occupations

Use as cited research evidence; verify terms per releaseUpdated research dashboard

Source ↗

How the monitor thinks

From exposure to real-world change.

A high structural score alone is not a forecast of unemployment. The monitor deliberately separates capability from adoption and realized labor-market outcomes.

01Structural exposure

How exposed are the occupation's tasks to current AI capability?

02Observed AI use

Is AI actually being used for those tasks in the real economy?

03Labor-market pressure

Are hiring, entry-level access, wages or skill premiums changing?

04Interpretation

Is the role being substituted, augmented, reorganized — or is the evidence still weak?

05Watch next

Which next data point would materially change the current picture?

No synthetic “job loss probability” will be shown unless the underlying evidence supports it. When evidence is incomplete, the product says so.