# Chandar and Klein Teeselink study AI and labor demand across 41 countries

[Canonical reading](https://escapemeridian.com/readings/chandar-and-klein-teeselink-study-ai-and-labor-demand-across-41-2ba355fc)

Bharat Chandar and Bouke Klein Teeselink’s working paper, originally dated September 20, 2026, and revised November 5, 2026, examines how generative AI adoption changes labor demand within firms. Their model separates a firm-wide productivity effect from “worker-type-specific changes in task content”—changes in the tasks different workers do. Using job-posting and employment records across 41 countries, the authors report that foreign affiliates of companies adopting AI reduce the junior share of their workforce relative to comparable control affiliates, mainly through growth in senior employment. They also report that senior employment shifts toward “AI-exposed occupations”—jobs with tasks AI can affect. Their point estimates suggest a shift away from those occupations among juniors, a result the authors present with uncertainty. The abstract describes changes in labor demand and occupational patterns; it does not establish how these patterns affect individual workers beyond the measures studied.

Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.

## Original source

Stanford Digital Economy Lab · Article

Original source date: Sep 20, 2026 (2026-09-20).

Date basis: Date recorded for this work at import.

Added to Escape Meridian: 2026-11-21T11:41:31.024Z. First observed: 2026-10-30T13:00:26.402Z. Retrieved: 2026-11-21T11:40:42.917Z.

- [How Does AI Change Labor Demand? Evidence from 41 Countries - Stanford Digital Economy Lab](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Limits

The evidence here is the publisher’s brief abstract for a working paper; it does not provide the instrument, event-study specifications, or estimates’ magnitudes. The page reports shifts in labor demand and occupational exposure but does not describe effects on individual workers beyond these measures.


## Open question

What worker-type-specific changes in task content underlie the point estimates suggesting a shift away from AI-exposed occupations among juniors?


## Source-stated

The authors study how generative AI changes labor demand among workers within a firm, using a model that separates firm-wide productivity effects from worker-type-specific changes in task content.

> Our model decomposes changes in labor demand following firm AI adoption into a firm-wide productivity effect and worker-type-specific changes in task content.

- [Supporting source](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Source-stated

The analysis uses 1.25 billion job postings and 154 million employment records across 41 countries.

> Using 1.25 billion job postings and 154 million employment records across 41 countries

- [Supporting source](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Source-stated

The authors report that foreign affiliates of AI-adopting companies reduce the junior share of their workforce relative to comparable control affiliates, mainly through growth in senior employment.

> foreign affiliates of AI-adopting companies reduce the junior share of their workforce relative to comparable control affiliates, mainly through growth in senior employment.

- [Supporting source](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Source-stated

The authors report that senior employment shifts toward AI-exposed occupations, while their point estimates suggest juniors shift away from those occupations.

> Senior employment shifts toward AI-exposed occupations, while our point estimates suggest a shift away from these occupations among juniors.

- [Supporting source](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Source-stated

The working paper’s original version is dated September 20, 2026, and it was revised November 5, 2026.

> Revised November 5, 2026, Original version September 20, 2026

- [Supporting source](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Source-stated

The authors are Bharat Chandar and Bouke Klein Teeselink.

> Authors
> Bharat Chandar, Bouke Klein Teeselink

- [Supporting source](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Source and contributor context

Stanford Digital Economy Lab · Publisher

The publisher is identified here. No individual author has been inferred.


## Source media

Media declared by the original publisher; availability checked separately from the editorial reading.

- [image · Stanford Digital Economy Lab](https://digitaleconomy.stanford.edu/app/uploads/2025/08/DEL_CoolNeutral_05-e1772819710191.png)

## Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.


## AI exposure

In labor economics, how much of a job's tasks an AI system could in principle perform; it is not a forecast of job loss.

- [Definition reference](https://arxiv.org/abs/2303.10130)

## Stanford Digital Economy Lab

Stanford Digital Economy Lab is a university research lab publishing studies on digital technology, employment and the economy.

- [Explore this context](/entities/stanford-digital-economy-lab)
- [Reference](https://digitaleconomy.stanford.edu/)

## Editorial provenance

Version 2. Publication ID: a123f4d3-f0d2-4e07-89f6-5908923ced8a.

Prepared by gpt-6-luna; grounding review by gpt-6-luna.

Source check: passed by anthropic/claude-opus-5.5 on 2026-11-21T11:41:31.024Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: material\_development.

Observation: 1ff30cc8647aa5e2ffe6cbb4ee33bf14892ec2cf89d96fea85a5f65f401e9869. Source text hash: 45c87c36d55eb53ec3daa9108615de046cd71b717e994eb6aee52182226cfaeb.

Correction reason: Added the paper’s original and revised dates, clarified the two requested terms, and attributed the publisher-abstract findings to Chandar and Klein Teeselink while retaining the uncertainty of the junior-shift estimates.


## Revision and source-check history

### Version 2 — 2026-11-21T11:41:31.024Z

Bharat Chandar and Bouke Klein Teeselink’s working paper, originally dated September 20, 2026, and revised November 5, 2026, examines how generative AI adoption changes labor demand within firms. Their model separates a firm-wide productivity effect from “worker-type-specific changes in task content”—changes in the tasks different workers do. Using job-posting and employment records across 41 countries, the authors report that foreign affiliates of companies adopting AI reduce the junior share of their workforce relative to comparable control affiliates, mainly through growth in senior employment. They also report that senior employment shifts toward “AI-exposed occupations”—jobs with tasks AI can affect. Their point estimates suggest a shift away from those occupations among juniors, a result the authors present with uncertainty. The abstract describes changes in labor demand and occupational patterns; it does not establish how these patterns affect individual workers beyond the measures studied.

## Original source

Stanford Digital Economy Lab · Article

Original source date: Sep 20, 2026 (2026-09-20).

Date basis: Date recorded for this work at import.

Added to Escape Meridian: 2026-11-21T11:41:31.024Z. First observed: 2026-10-30T13:00:26.402Z. Retrieved: 2026-11-21T11:40:42.917Z.

- [How Does AI Change Labor Demand? Evidence from 41 Countries - Stanford Digital Economy Lab](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Limits

The evidence here is the publisher’s brief abstract for a working paper; it does not provide the instrument, event-study specifications, or estimates’ magnitudes. The page reports shifts in labor demand and occupational exposure but does not describe effects on individual workers beyond these measures.


## Open question

What worker-type-specific changes in task content underlie the point estimates suggesting a shift away from AI-exposed occupations among juniors?


## Source-stated

The authors study how generative AI changes labor demand among workers within a firm, using a model that separates firm-wide productivity effects from worker-type-specific changes in task content.

> Our model decomposes changes in labor demand following firm AI adoption into a firm-wide productivity effect and worker-type-specific changes in task content.

- [Supporting source](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Source-stated

The analysis uses 1.25 billion job postings and 154 million employment records across 41 countries.

> Using 1.25 billion job postings and 154 million employment records across 41 countries

- [Supporting source](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Source-stated

The authors report that foreign affiliates of AI-adopting companies reduce the junior share of their workforce relative to comparable control affiliates, mainly through growth in senior employment.

> foreign affiliates of AI-adopting companies reduce the junior share of their workforce relative to comparable control affiliates, mainly through growth in senior employment.

- [Supporting source](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Source-stated

The authors report that senior employment shifts toward AI-exposed occupations, while their point estimates suggest juniors shift away from those occupations.

> Senior employment shifts toward AI-exposed occupations, while our point estimates suggest a shift away from these occupations among juniors.

- [Supporting source](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Source-stated

The working paper’s original version is dated September 20, 2026, and it was revised November 5, 2026.

> Revised November 5, 2026, Original version September 20, 2026

- [Supporting source](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Source-stated

The authors are Bharat Chandar and Bouke Klein Teeselink.

> Authors
> Bharat Chandar, Bouke Klein Teeselink

- [Supporting source](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Source and contributor context

Stanford Digital Economy Lab · Publisher

The publisher is identified here. No individual author has been inferred.


## Source media

Media declared by the original publisher; availability checked separately from the editorial reading.

- [image · Stanford Digital Economy Lab](https://digitaleconomy.stanford.edu/app/uploads/2025/08/DEL_CoolNeutral_05-e1772819710191.png)

## Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.


## AI exposure

In labor economics, how much of a job's tasks an AI system could in principle perform; it is not a forecast of job loss.

- [Definition reference](https://arxiv.org/abs/2303.10130)

## Stanford Digital Economy Lab

Stanford Digital Economy Lab is a university research lab publishing studies on digital technology, employment and the economy.

- [Explore this context](/entities/stanford-digital-economy-lab)
- [Reference](https://digitaleconomy.stanford.edu/)

## Editorial provenance

Version 2. Publication ID: a123f4d3-f0d2-4e07-89f6-5908923ced8a.

Prepared by gpt-6-luna; grounding review by gpt-6-luna.

Source check: passed by anthropic/claude-opus-5.5 on 2026-11-21T11:41:31.024Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: material\_development.

Observation: 1ff30cc8647aa5e2ffe6cbb4ee33bf14892ec2cf89d96fea85a5f65f401e9869. Source text hash: 45c87c36d55eb53ec3daa9108615de046cd71b717e994eb6aee52182226cfaeb.

Correction reason: Added the paper’s original and revised dates, clarified the two requested terms, and attributed the publisher-abstract findings to Chandar and Klein Teeselink while retaining the uncertainty of the junior-shift estimates.


### Version 1 — 2026-11-21T10:52:01.897Z

Bharat Chandar and Bouke Klein Teeselink study how generative AI adoption changes labor demand among workers within a firm. Their model separates a firm-wide productivity effect from worker-type-specific changes in task content. Using 1.25 billion job postings and 154 million employment records across 41 countries, they report that foreign affiliates of companies adopting AI reduce the junior share of their workforce relative to comparable control affiliates, mainly through growth in senior employment. Senior employment shifts toward AI-exposed occupations, while their point estimates suggest juniors shift away from those occupations. The authors’ findings describe different employment patterns for senior and junior workers; the brief publisher-page text does not establish how these patterns affect workers beyond the measured labor demand.

## Original source

Stanford Digital Economy Lab · Article

Original source date: Publication date not provided.

Date basis: Date basis not provided.

Added to Escape Meridian: 2026-11-21T10:52:01.897Z. First observed: 2026-10-30T13:00:26.402Z. Retrieved: 2026-11-21T10:48:36.947Z.

- [How Does AI Change Labor Demand? Evidence from 41 Countries - Stanford Digital Economy Lab](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Limits

The evidence is a brief publisher-page description of a working paper; it does not provide the instrument, event-study specifications, or estimates’ magnitudes. The page reports a shift in labor demand and occupational exposure, but does not describe effects on individual workers beyond these measures.


## Open question

What worker-type-specific changes in task content underlie the reported shift away from AI-exposed occupations among juniors?


## Source-stated

The authors study how generative AI changes labor demand among workers within a firm, using a model that separates firm-wide productivity effects from worker-type-specific changes in task content.

> Our model decomposes changes in labor demand following firm AI adoption into a firm-wide productivity effect and worker-type-specific changes in task content.

- [Supporting source](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Source-stated

The analysis uses 1.25 billion job postings and 154 million employment records across 41 countries.

> Using 1.25 billion job postings and 154 million employment records across 41 countries

- [Supporting source](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Source-stated

The authors report that foreign affiliates of AI-adopting companies reduce the junior share of their workforce relative to comparable control affiliates, mainly through growth in senior employment.

> foreign affiliates of AI-adopting companies reduce the junior share of their workforce relative to comparable control affiliates, mainly through growth in senior employment.

- [Supporting source](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Source-stated

The authors report that senior employment shifts toward AI-exposed occupations, while their point estimates suggest juniors shift away from those occupations.

> Senior employment shifts toward AI-exposed occupations, while our point estimates suggest a shift away from these occupations among juniors.

- [Supporting source](https://digitaleconomy.stanford.edu/publication/how-does-ai-change-labor-demand/)

## Source and contributor context

Stanford Digital Economy Lab · Publisher

The publisher is identified here. No individual author has been inferred.


## Source media

Media declared by the original publisher; availability checked separately from the editorial reading.

- [image · Stanford Digital Economy Lab](https://digitaleconomy.stanford.edu/app/uploads/2025/08/DEL_CoolNeutral_05-e1772819710191.png)

## Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.


## AI exposure

In labor economics, how much of a job's tasks an AI system could in principle perform; it is not a forecast of job loss.

- [Definition reference](https://arxiv.org/abs/2303.10130)

## Stanford Digital Economy Lab

Stanford Digital Economy Lab is a university research lab publishing studies on digital technology, employment and the economy.

- [Explore this context](/entities/stanford-digital-economy-lab)
- [Reference](https://digitaleconomy.stanford.edu/)

## Editorial provenance

Version 1. Publication ID: 57638146-eb8e-4d5b-80bc-09ffc626d6f9.

Prepared by gpt-6-luna; grounding review by gpt-6-luna.

Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-21T10:52:01.897Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: material\_development.

Observation: 1ff30cc8647aa5e2ffe6cbb4ee33bf14892ec2cf89d96fea85a5f65f401e9869. Source text hash: 45c87c36d55eb53ec3daa9108615de046cd71b717e994eb6aee52182226cfaeb.


### uncertain — 2026-11-21T10:52:01.897Z — anthropic/claude-opus-5.5

Publication: 57638146-eb8e-4d5b-80bc-09ffc626d6f9

attribution — pass: The draft correctly credits the working paper to Bharat Chandar and Bouke Klein Teeselink, published by the Stanford Digital Economy Lab. All findings are framed as the authors' reports, not as established fact.

> Authors
Bharat Chandar, Bouke Klein Teeselink

summary — uncertain: The summary follows the source faithfully, but the question calls the juniors' move a 'reported shift away'. The source says only that point estimates 'suggest' it, which is weaker than the senior finding. Smallest repair: change it to 'suggested shift away' and keep that hedge visible. Central terms such as 'point estimates', 'AI-exposed occupations' and 'foreign affiliates' versus control affiliates go unexplained in plain language; tooltips are needed. The summary is also thin on why the finding matters, though it honestly states what remains unknown.

> while our point estimates suggest a shift away from these occupations among juniors

date — uncertain: evidence.originalPublished is null with unknown precision, so no original date would be displayed. The source text gives an original version dated September 20, 2026, a revision on November 5, 2026, and a listed date of 09/21/2026. The first-publication date is therefore available but not captured. It should be September 20, 2026, not the later revision date.

> Revised November 5, 2026, Original version September 20, 2026

support — pass: Every claim matches verbatim source text and is presented as source-stated. The limitations accurately note what the brief page omits. No benefits, costs or forecasts are invented.

> Using 1.25 billion job postings and 154 million employment records across 41 countries

media — pass: The selected media is the verified direct link to the original publisher page. It is the most direct way for readers to reach the source.

### passed — 2026-11-21T11:41:31.024Z — anthropic/claude-opus-5.5

Publication: a123f4d3-f0d2-4e07-89f6-5908923ced8a

attribution — pass: The reading correctly names Bharat Chandar and Bouke Klein Teeselink as the authors. It presents the findings as the authors' reported results, not as established fact, and identifies the source as a working paper.

> Authors
Bharat Chandar, Bouke Klein Teeselink

summary — pass: The reading faithfully conveys the model's decomposition, the data scale, the junior-share result driven by senior growth, and the hedged junior occupational shift. Central terms such as task content and AI-exposed occupations get plain glosses. The term 'point estimates' could use a term-library tooltip, but the reading signals the hedge. The title and question follow directly from the evidence.

> while our point estimates suggest a shift away from these occupations among juniors

date — pass: The displayed originalPublished date, 2026-09-20, matches the stated original version date. The November 5 revision is correctly treated as a later revision. The page also shows a separate 09/21/2026 date, but the explicit original-version label supports September 20.

> Revised November 5, 2026, Original version September 20, 2026

support — pass: Each claim is directly supported by the abstract. The gloss of AI-exposed occupations as 'jobs with tasks AI can affect' is a light plain-language explanation, not a new factual claim. The limitations accurately note that the instrument details and effect magnitudes are absent, and the reading does not imply unsupported benefits or harms.

> foreign affiliates of AI-adopting companies reduce the junior share of their workforce relative to comparable control affiliates, mainly through growth in senior employment

media — pass: The selected media is the verified direct link to the original publisher page for this working paper. It is the most relevant way for readers to access the source.

## Recorded references

Explicit citations in public readings to this reading. Shared topics are not citations. Other record types need an explicit reference adapter.

No public references recorded here yet.

[Reference data (JSON)](/readings/2ba355fc4cc89444bf0075a9e719fde6d63caad1594a003abda311f278db2638/references.json)

## Corrections and challenges

[Send a correction or an opposing source](https://escapemeridian.com/corrections/)

[JSON record](https://escapemeridian.com/readings/chandar-and-klein-teeselink-study-ai-and-labor-demand-across-41-2ba355fc.json)
