---
id: "lib-2026-08-12-canaries-august-release-note"
type: "library"
title: "Canaries August release note"
lang: "EN"
org: "Stanford DEL"
source_type: "Article"
date: "2026-08-12"
week: "2026-W33"
url: "https://nosov.ai/library/lib-2026-08-12-canaries-august-release-note/"
original_url: "https://digitaleconomy.stanford.edu/news/canariesaug26/"
tier: "B"
tier_name: "Useful, limited"
tier_reason: "A short release note by Brynjolfsson, Chandar and Chen summarising their own revised paper; reliable on figures but a summary, with the analysis in the paper itself."
summary_basis: "full-text"
original_minutes: "7"
tags: ["labor market"]
---

# Canaries August release note

Tier B (Useful, limited): A short release note by Brynjolfsson, Chandar and Chen summarising their own revised paper; reliable on figures but a summary, with the analysis in the paper itself.

Summary based on the full text.

## In short

A Stanford Digital Economy Lab post (August 2026) by Erik Brynjolfsson, Bharat Chandar and Ruyu Chen announcing the revised "Canaries in the Coal Mine?" paper. Using ADP payroll data through June 2026, they find no economy-wide AI job displacement, but the employment gap for young workers in AI-exposed occupations has widened from 15% to 19%. The note also announces a monthly Canaries Dashboard inside the lab's new AI Economic Indicators.

## Key numbers

- 19%: employment shortfall for workers aged 22-25 in highly AI-exposed occupations versus less-exposed peers, June 2026
- 15%: the same shortfall at the July 2025 data vintage
- -11% / +10%: employment change, Nov 2022 to Jun 2026, for 22-25-year-olds in the two most exposed vs the three least exposed quintiles

## What it reports

- No widespread, economy-wide job displacement associated with AI.
- Employment of workers aged 22-25 in highly AI-exposed occupations is about 19% below where it would be if it had kept pace with less-exposed peers of the same age; experienced workers show no comparable gap.
- The shortfall was 15% at the July 2025 data vintage and 19% as of June 2026.
- In levels: employment of 22-25-year-olds in the two most exposed quintiles fell about 11% between November 2022 and June 2026, while the three least exposed quintiles grew about 10%.
- Adjustment runs mainly through reduced hiring of young workers, not more separations, and shows up in employment rather than base pay (which excludes bonuses and equity).
- Declines concentrate where AI usage automates tasks; where usage complements workers, employment is flat or rising, especially for experienced staff.
- New mechanism evidence: young-worker employment fell in occupations built on codified knowledge, while experienced-worker employment rose in occupations built on tacit knowledge. Women face greater AI exposure on average.

## How they got it

Descriptive analysis of ADP administrative payroll data, comparing employment trends by age group and occupational AI exposure, using exposure measures from Eloundou et al. (2024) and usage data from the Anthropic Economic Index. Robustness checks exclude technology firms and computer occupations, control for interest-rate exposure and remote work, and allow firm entry and exit.

## Limits

- The authors call these descriptive patterns, not causal estimates.
- Gaps shrink when controlling for education; some divergent trends predate generative AI; gaps are larger in the ADP sample than in national survey benchmarks, with the discrepancy concentrated in education, health care and public administration.
- Firm-adjusted estimates became more sensitive to specification after pipeline changes.

## Why it matters for companies

The effect of AI on staffing is showing up first as fewer entry-level hires in codified-knowledge roles, not as layoffs. A company that cuts graduate intake in such roles saves cost now but thins the bench that builds tacit expertise later. Aggregate headcount numbers will not reveal this; managers need to track hiring by age band and role type. The monthly dashboard gives HR a public benchmark to compare against internal data.

Original: https://digitaleconomy.stanford.edu/news/canariesaug26/