aijobimpactcalculator.com
Menu
CalculatorOccupationsIndustriesWhat's growingMethodologyFAQ

Reference / Sources / Last verified June 2026

Sources: the full bibliography

Every primary source the calculator draws on, classified by what it publishes (per-occupation, per-task, or aggregate) and how the calculator uses it. The page is structured to be cited entry by entry.

Per-occupation primary sources

PER OCCUPATION / 2025

Generative AI and Jobs: A Refined Global Index of Occupational Exposure (ILO Working Paper 140)

Gmyrek, Berg, Kaminski et al. — International Labour Organization & NASK (published 20 May 2025)

Scope. ISCO-08 6-digit occupations across approximately 30,000 tasks and more than 50,000 human assessments; four exposure gradients per occupation.

Use here. Primary source for the calculator's exposure-gradient output. Cross-referenced with OECD AI Working Group output for triangulation.

Caveat. Uses ISCO-08 codes. Mapping to O*NET-SOC is via the BLS-published SOC-to-ISCO crosswalk, which is one-to-many in some cases; the calculator uses the dominant match.

Accessed: 2026-04

https://www.ilo.org/publications/generative-ai-and-jobs-2025-update

PER OCCUPATION / 2026

Employment Projections 2025-2035

US Bureau of Labor Statistics

Scope. Decade employment projections for approximately 800 detailed occupations and 300 industries through 2035.

Use here. Primary source for the calculator's growth panel (per-occupation projected employment change). Figures taken from the National Employment Matrix (data.bls.gov/projections).

Caveat. BLS projections do not explicitly model AI as a discrete disruption variable; they reflect aggregate workforce dynamics including demographics, demand, and policy.

Accessed: 2026-09

https://www.bls.gov/news.release/ecopro.htm

PER OCCUPATION / 2025

Occupational Outlook Handbook

US Bureau of Labor Statistics

Scope. Per-occupation employment outlook narratives, projected growth, median wage, and entry requirements.

Use here. Reference source for outlook narratives on per-occupation deep-dive pages.

Accessed: 2026-04

https://www.bls.gov/ooh/

PER OCCUPATION / 2025

O*NET 30.2 Database

O*NET Resource Center / National Center for O*NET Development

Scope. Approximately 1,000 occupations, 19,000 task statements, plus skills, knowledge, abilities, work activities.

Use here. Source for the occupation taxonomy, top-five-task list per occupation, similar-occupation links, and skills profile.

License: CC-BY 4.0 (attribution required).

Accessed: 2026-04

https://www.onetcenter.org/database.html

Per-task primary sources

PER TASK / 2024

Generative AI, the American Worker, and the Future of Work

Brookings Institution

Scope. Task-level automation exposure across O*NET tasks for over 1,000 occupations.

Use here. Source for the task-level Displaceable / Changing / Growing tags applied to each occupation's top five O*NET tasks.

Caveat. Brookings's task-level data is built on OpenAI's task-completion likelihood data. Brookings flags this; this calculator notes the same limitation on the methodology page.

Accessed: 2026-04

https://www.brookings.edu/articles/generative-ai-the-american-worker-and-the-future-of-work/

Aggregate sources (context only)

AGGREGATE / 2025

Future of Jobs Report 2025

World Economic Forum

Scope. Global skills outlook plus growing/declining occupation lists at the family level (170M new roles, 92M displaced, net 78M new by 2030).

Use here. Source for the calculator's growing-skills panel.

Caveat. WEF data is family-level, not per-occupation. The calculator uses it for the skills panel only.

Accessed: 2026-04

https://www.weforum.org/publications/the-future-of-jobs-report-2025/

AGGREGATE / 2024

A New Future of Work: The Race to Deploy AI and Raise Skills in Europe and Beyond

McKinsey Global Institute

Scope. Aggregate workforce-impact estimates: 30% of current hours could be automated by 2030 (midpoint), 50% by 2030-2060.

Use here. Reference only on the homepage and methodology page for time-horizon framing.

Caveat. Commercial research firm. Cited for framing, not as primary methodology, to avoid consultant-for-the-disruptor optics.

Accessed: 2026-04

https://www.mckinsey.com/mgi/our-research/a-new-future-of-work-the-race-to-deploy-ai-and-raise-skills-in-europe-and-beyond

Reference and policy framing

Excluded sources, with reasons

EXCLUDED / 2013

The Future of Employment: How Susceptible Are Jobs to Computerisation?

Frey and Osborne

Scope. Probability-of-computerisation classifier across 702 occupations, derived pre-LLM.

Use here. Excluded as primary methodology because the underlying probability estimates are pre-LLM and miss the cognitive-task disruption that defines the 2024-2026 shift. Acknowledged as the foundational paper that powered earlier-generation calculators.

Accessed: 2026-04

https://www.oxfordmartin.ox.ac.uk/downloads/academic/The_Future_of_Employment.pdf

EXCLUDED / 2023

GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models

Eloundou, Manning, Mishkin, Rock

Scope. Task-level exposure analysis using LLM self-report and human annotators.

Use here. Excluded as primary methodology because the analysis is OpenAI-derived; using a frontier-AI vendor's own work as primary methodology for an AI-displacement calculator carries optics risk. Cited for completeness on the sources page.

Accessed: 2026-04

https://arxiv.org/abs/2303.10130

Cross-references: see the full methodology at /methodology/, the pre-empted challenges at /how-to-argue-with-this/, and the term definitions at /glossary/.

Questions

Didn't find your answer?

Ask us. A real person reads every question and we answer the ones we can, with sources. If your question would help other readers, we may publish an anonymised version, with your permission. General reference only.

Ask a question

questions@aijobimpactcalculator.com