AI use and occupationsFatih Kansoy ↗
Occupation observations April–May 2026Five observation windows Definitions
The time dimension

Five windows. Two monthly occupation observations.

Observation dates describe the activity in the source. Release dates describe when the publisher made it available. Neither is the date this website was updated.

Why only April and May?

The pinned June 2026 release publishes country-level occupation shares for April and May. The selected earlier releases contain country task snapshots from August 2025, November 2025 and February 2026. They remain available in the data catalogue, but they have different windows and measurement procedures. A line connecting them would suggest a comparable series that these files do not establish.

The official source repository was checked on 9 September 2026; the latest available release remained 26 June. Missing months have not been interpolated.

Publication coverage by window

4–11 August 2025

Released 15 September 2025

113 countries and areas with published task evidence

Weekly task snapshot
13–20 November 2025

Released 15 January 2026

116 countries and areas with published task evidence

Weekly task snapshot
5–12 February 2026

Released 24 March 2026

117 countries and areas with published task evidence

Weekly task snapshot
April 2026

Released 26 June 2026

114 countries and areas with published task evidence

Monthly task + occupation facets
May 2026

Released 26 June 2026

121 countries and areas with published task evidence

Monthly task + occupation facets

Inspect the historical task windows

Bars show published task mass within each separate window. Changes can reflect sampling, classifiers, disclosure and taxonomy as well as behaviour.

Country counts and the registry

The headline is 121 countries and areas with May occupation observations: 118 UN member states, the State of Palestine, Puerto Rico and Taiwan. The 250-entry directory is a lookup registry: 248 UNSD M49 entries plus Kosovo and Taiwan. Its unobserved entries are useful for identifying gaps, but they are not measured observations.

Employment context is available for 182 registry entries; 113 intersect the May occupation data. Availability alone does not make their statistical populations, years or classifications interchangeable.

Pinned primary source · Source observation inventory · Construction documentation