Data Downloads

Name: AI Work Index — Structural Scores and Country Bundles
Creator: AI Work Index
Published: 2026-06-11
License: https://opensource.org/licenses/MIT

The global structural baseline and the country-specific bundles are published separately so each layer can be validated on its own, without conflating structural pressure with local demand, wage, transition, or regulatory overlays.

562 occupations · 88 roles · 57 data sources · MIT licensed

Structural scores and local context are separate downloads. Each artifact has an evidence tier: official local, derived from official local, cross-country research, external proxy, or synthetic.

Forecast readiness is published separately

4 inputs are ready for directional validation, 2 need recurring snapshots, and 2 are source-available but not yet modelled. The matrix does not change the V7 headline score.

forecast-readiness-v7.json

Structural release

score dataset generated 2026-06-11

Live monitor

Q4 2025 full

current labour context used on live pages

Latest official release

Q4 2025

MOM Labour Market Report Q4 2025 published 2026-03-20

Legacy model comparison

Shadow published

median direct task-share 100%

Quarterly briefing

2026 Q1

occupations-v7-2026-06.json

Release history

V7 structural release

Published 7 Apr 2026 · V7 · monitor Q4 2025 full

Open →

V6 structural release

Published 1 Apr 2026 · V6 · monitor Q4 2025 full

Open →

V5 structural release

Published 21 Mar 2026 · V5 · monitor Q4 2025 full

Open →

View 14 older releases

V4.3 structural release

Published 21 Mar 2026 · V4.3 · monitor Q4 2025 full

Open →

V4.2 structural release

Published 21 Mar 2026 · V4.2 · monitor Q4 2025 full

Open →

V4.0 four-source ensemble milestone

Published 19 Mar 2026 · V4.0 · monitor Q4 2025 full

Open →

V3.3 ensemble exposure iteration

Published 19 Mar 2026 · V3.3 · monitor Q4 2025 full

Open →

V3.2 confidence-interval milestone

Published 19 Mar 2026 · V3.2 · monitor Q4 2025 full

Open →

V3.1 observed-exposure and demand-signal milestone

Published 16 Mar 2026 · V3.1 · monitor Q4 2025 full

Open →

V3.0 three-layer structural score

Published 16 Mar 2026 · V3.0 · monitor Q4 2025 full

Open →

V2 Singapore occupation scorer

Published Jan 2026 · V2 · monitor Q4 2025 full

Open →

V1 public alpha

Published Dec 2025 · V1 · monitor Q4 2025 full

Open →

V4.3 shadow score published

Published 21 Mar 2026 · V7 · monitor Q4 2025 full

Open →

V5 sidecars published

Published 21 Mar 2026 · V7 · monitor Q4 2025 full

Open →

V5 experimental model published

Published 21 Mar 2026 · V7 · monitor Q4 2025 full

Open →

MOM Labour Market Report Q4 2025

Published 20 Mar 2026 · V7 · monitor Q4 2025 full

Open →

2026 Q1 quarterly briefing

Published 21 Mar 2026 · V7 · monitor Q4 2025 full

Open →

Downloads

CSV

562 occupations, flattened fields + provenance. Best for spreadsheets.

sg-ai-occupations-v7.csv

JSON

Full V7 scores with nested fields, task primitives, uncertainty intervals, and retained V6 baseline fields.

sg-ai-occupations-v7.json

V4.3 Audit Trail

Shadow-governance status, promotion gates, and the retained audit trail beneath the live V7 release.

experimental-methodology-v43.json

V4.3 Comparison Scores

Task-adjusted comparison scores retained for archived validation and promotion-gate review.

shadow-scores-v43.json

Research Library

Canonical citation registry linking the live methodology, validation layer, and archived V5 roadmap to source papers and reports.

research-library.json

V5 Roadmap

Archived roadmap for the V5 research program that preceded the current V7 two-axis structural release.

v5-roadmap.json

V5 Supplementary Data

Published V5 workstream summary covering augmentation heterogeneity, empirical mobility, posterior uncertainty, and realized-risk forecasting sidecars that fed the later V7 release.

v5-sidecars.json

V5 Model Note

Final promotion-comparison artifact for the former live V5 model, retaining the V4.3 comparison snapshot and the published adjunct layers.

v5-experimental-model.json

V5 Validation Comparison

Comparison and validation summary for the former live V5 model versus the retained V4.3 snapshot across structural and realized-risk checks.

v5-experimental-validation.json

Forecast Readiness

Source, duplication, and validation-gate matrix for future forecast-grade labour claims. Separate from the live structural score.

forecast-readiness-v7.json

Source Code

Full scoring pipeline and raw data. Open source, MIT licensed.

github.com/kirso/aiworkindex

Country Context Bundles

Context-only bundles are published separately from structural scores. Singapore is fully scored, while the US layer now carries its own public wage and demand context.

Full context pack

All context in one file

Labour monitor

Vacancy, hiring, retrenchment

Worker profile

Age, education, nationality

Geography

Planning area concentration

Macro context

Unemployment, GDP, tightness

AI in Singapore

Adoption, NAIIP, workforce

O*NET task + tools

Supporting task and technology context

Transition support

Pathways, SkillsFuture, JTM / WSQ anchors

Offset potential

Demand persistence, redesign room, switching friction

Forecast readiness

Sources, gaps, validation gates

Field source map

Field-level provenance and transformations

Versioned Snapshots

Historical scoring snapshots for tracking changes over time.

Version	Date	Occupations	Download
V7 (Current)	7 Apr 2026	562	JSON · CSV
V6	1 Apr 2026	562	Archived
V5	21 Mar 2026	562	Archived
V4.3	21 Mar 2026	562	Archived
V4.2	21 Mar 2026	562	JSON · CSV
V4.0	19 Mar 2026	562	History only
V3.3	19 Mar 2026	562	History only
V3.2	19 Mar 2026	562	History only
V3.1	16 Mar 2026	562	History only
V3.0	16 Mar 2026	562	JSON
V2	Jan 2026	562	History only
V1	Dec 2025	562	History only

Quarterly snapshots will be archived here as new scoring runs are published. Use snapshots to track band movers and score drift over time.

Methodology Version

Version: V7 (headline risk = displacement pressure × (1 − demand resilience), with V7 task-concentration-weighted exposure and demand-persistence proxy)

Data vintage: 2024 wages, 2025 labour context, 2025/2026 demand signals

Occupations: 562 occupations, currently sourced from Singapore (primary) and United States data

Separate context bundle: Labour monitor, worker profile, industry context, sector wage anchors, geography context, macro labour context, national AI context, offset potential, transition support, and US wage / requirements / skills / narrative layers

Retained baseline trail: Shadow published. The full V4.3 shadow and V5 promotion comparison remain published so the live V7 release can still be audited against the retained V4.3 and V4.2 baselines.

Research memory: 34 canonical research entries are published in the research library and linked to claims/source registry records.

Sources: Live reference-market official statistics and policy data (wages, labour-force context, industry context, demand signals), MOM AI Adoption Among Firms 2026, BLS OEWS, ORS, CPS demographics, Skills Data, OOH, IMDA Singapore Digital Economy Report 2025, IMDA NAIIP 2026, O*NET, Felten AIOE, Pizzinelli/IMF, Anthropic observed usage, Anthropic labor-market impacts, Eloundou GPT exposure, ILO occupational exposure, SOL 2026, Jobs in Demand 2025

Full methodology → V7 release note → Research library →

Release Metadata (54 artifacts)

Manifest version: V7

Manifest generated: 12 Jun 2026, 7:18 pm

Score dataset vintage: 2026-06-11

Artifact	Generated	Size	SHA-256
V7 structural score CSV sg-ai-occupations-v7.csv	12 Jun 2026, 5:53 pm	345.0 KB	edd48803b1ead2a3f952cdc33ca86be7c50419382fe59691f0c65b5738e54052
V7 structural score JSON sg-ai-occupations-v7.json	12 Jun 2026, 5:53 pm	3.28 MB	8caadf0d560a4b6ace5ae5bb9e531ab6b2a790569cf815e5a3f5b05b89eb6f77
V4.3 structural score CSV snapshot sg-ai-occupations-v43.csv	15 May 2026, 4:04 pm	355.0 KB	a2cf27e379fb9cd9afdb84868a51d12bcc6e0aaa5648776af612dca9073a86a9
V4.3 structural score JSON snapshot sg-ai-occupations-v43.json	11 Jun 2026, 10:36 pm	4.03 MB	0cd2872762d6f20bedbfecce260538a4e2249dc32cae97c5fba74befafa6450b
V5 structural score CSV snapshot sg-ai-occupations-v5.csv	15 May 2026, 4:04 pm	367.9 KB	41d9f2678bd959bfbbd6d6cfd0aafb9c4a63ba601ae339413270fab7c089411a
V5 structural score JSON snapshot sg-ai-occupations-v5.json	11 Jun 2026, 10:36 pm	4.12 MB	2d9716694eecb86dc1ec3cbb1a508061a3b592b305af7ffe4b71b2f3a1a73621
V4.2 structural score CSV snapshot sg-ai-occupations-v42.csv	15 May 2026, 4:04 pm	268.0 KB	8faad00e56e86503d75762fbde041145d1108552ec7fc5a4ef48867d369b45cf
V4.2 structural score JSON snapshot sg-ai-occupations-v42.json	11 Jun 2026, 10:36 pm	3.02 MB	63029c9709c7b841a694107798d4396428108963120b2f8b75d72b5093b28269
O*NET task and technology enrichment onet-enrichment.json	7 Jun 2026, 7:57 pm	1.33 MB	cc2698082ca62fa3c336815590c1a496f8f0db025f4a49efceedfa5bbff5200f
Singapore context pack sg-context-pack-2025.json	12 Jun 2026, 5:53 pm	655.8 KB	69384d017841becdfcbfb98e33c1d6a8f044c3477c0482fcd12c7be78e80190a
Singapore labour monitor sg-labour-monitor-2025.json	11 Jun 2026, 10:36 pm	27.2 KB	cfacd54ba28bc34cb6e9feb32eb09b04e80de669d0056e9ffaa167cd92f9ab13
Singapore worker profile sg-worker-profile-2025.json	7 Jun 2026, 7:57 pm	114.5 KB	50dac9fb85e2b7decb34cbf8420d47dc250927da13e654fca9a24647467c0301
Singapore Section D deltas sg-lfr-deltas-2025.json	11 Jun 2026, 10:36 pm	43.6 KB	9642f9b7f5c7a0064389c9643f56b28c9326cde85dda5c5178ec358a4034c70c
Singapore geography context sg-geography-context-2020.json	12 Jun 2026, 5:53 pm	12.1 KB	7333ab0eb9976abc03f1a0df70a966afbb32375eb234f0b312904f4eb1f6513d
Singapore macro labour context sg-macro-context-2025.json	12 Jun 2026, 5:53 pm	20.4 KB	ed56992d4e66fec3f0431829a49f8b5d9159b53b3012a219e904c78c6994f663
AI in Singapore context sg-ai-in-singapore-2025.json	12 Jun 2026, 5:53 pm	5.8 KB	c69b92c5c41428e0c45f8b7de1c4a263de54c5334e4603f95272bcdfe4d4eac6
Transition infrastructure layer sg-transition-infrastructure-2025.json	12 Jun 2026, 5:53 pm	5.3 KB	ffc2d0b735925c7d27c5d7d419b82f28d1c2073269f0fa92f2d079cac60abea8
Transition support layer sg-transition-support-v4.json	12 Jun 2026, 5:53 pm	7.32 MB	204023ef0a6c4913c602e9b63e6a8abd9ceb8dac6f8870c1b5c8491208549577
United States evidence support bundle countries/us/support.json	13 Apr 2026, 11:21 am	6.84 MB	b390fdc59479d7ea6ade3cebea18c32b8c2fa74d5bbda9345d598ef1bcd0ccd5
Offset potential layer sg-offset-potential-v4.json	12 Jun 2026, 5:53 pm	521.0 KB	5969ea39b5aea75b0a1ff0268fc3b981e6fe74d1fd06ec0f8b536ae497b4cc77
Public claims matrix claims-matrix-v7.json	12 Jun 2026, 5:53 pm	13.8 KB	546783d6f970c596b562198fffd9d0675844b80a7a8c037486f3b130df7efbf2
Public field source map public-field-source-map.json	12 Jun 2026, 5:53 pm	12.5 KB	2bc38e9e7a10bdfb09a4a64600077d8b79a9991325253ee036d167ed4ef2db0e
Forecast readiness matrix forecast-readiness-v7.json	12 Jun 2026, 5:53 pm	19.7 KB	f06e61f88a6781166d3023ba55e0b84775bc0e5dff6b96dea0e445ba7d43958e
V4.3 shadow-model readiness experimental-methodology-v43.json	12 Jun 2026, 5:53 pm	3.4 KB	f08b6a79268e1aa4426feeada255ad6673a9e53c5d860e96445fad098b1a4e2b
V4.3 shadow scores shadow-scores-v43.json	12 Jun 2026, 5:53 pm	458.3 KB	cf5b40e4668d5e33c55d44b66f49f21a5d4fb272e2e6e74f32b43c71c5bdb4c6
V4.3 shadow comparison summary shadow-comparison-v43.json	12 Jun 2026, 5:53 pm	45.1 KB	d1ea7810fbfec3a7effbcef10a21a4caaf42e523f28b1435df8d8bc71ee3afe8
V4.3 shadow validation comparison shadow-validation-v43.json	12 Jun 2026, 5:53 pm	1.7 KB	4a1a32dc40524bd5a9bf67d23368d0ec077070842b5657f4f1028e17329334a0
V4.3 shadow anchor review shadow-anchor-review-v43.json	12 Jun 2026, 5:53 pm	4.3 KB	194139e99fbaa601f73ee99b74a9c4e8b0b36e1d7e4eb3b7b93d97d5af20fe0f
Research library research-library.json	12 Jun 2026, 5:53 pm	39.9 KB	ceeb12cbb0ff9f95cf371ddbf9dc4971280158da139adb621fbcceb646e51fcf
V5 roadmap v5-roadmap.json	12 Jun 2026, 5:53 pm	6.6 KB	9fa138f0574ee3ffcb0e340b6105210ffd33d3bf100102edf66cac2e162400c2
V5 sidecar summary v5-sidecars.json	12 Jun 2026, 5:53 pm	2.6 KB	e0d8800531913841dab1bdc4106800030d8d7157f84d9916ed855171cd02257a
V5 augmentation heterogeneity sidecar v5-augmentation-heterogeneity.json	12 Jun 2026, 5:53 pm	302.3 KB	388b2cddfe29aeee46fbc77b45e6ae61a51f727635a21af34ad38f3a982f0cfe
V5 empirical mobility sidecar v5-empirical-mobility.json	12 Jun 2026, 5:53 pm	513.0 KB	f4da6ca0841c0f6f3da655a2ca688ad16edfbdf9448ad0021653d4d37f51bef5
V5 posterior uncertainty sidecar v5-posterior-uncertainty.json	12 Jun 2026, 5:53 pm	610.2 KB	a088c17bb336a35939f403c5f7236e3935660a1db871b1043729145c1a189748
V5 realized-risk sidecar v5-realized-risk.json	12 Jun 2026, 5:53 pm	791.5 KB	0d6ca93f04896b67aa685149b809d9e43b062facbcfb6973d46ae0355cea3372
V5 experimental model v5-experimental-model.json	12 Jun 2026, 5:53 pm	1.67 MB	3e58eb91c1e38e41bacef52c35087fc1bcbb7e2b28b4c492eddadbe04e678166
V5 experimental validation v5-experimental-validation.json	12 Jun 2026, 5:53 pm	6.2 KB	c1af2021edf60f33a128141b1288aadb57d4db574a9d2406ed5d421ca2e21719
Quarterly report quarterly-report.json	12 Jun 2026, 5:53 pm	8.2 KB	eecb50109bd798491d19cbc56e2f84983e936bd0ba060a75987373f5e1b62b80
LLMs summary ../llms.txt	12 Jun 2026, 5:53 pm	2.1 KB	e2b28be42181c931e9825793047610ac661b39f9e913823254ee5a8f8ed15482
LLMs full reference ../llms-full.txt	12 Jun 2026, 5:53 pm	15.6 KB	e5dda2ab06f3cf62cb92896e371958f28a95f7bc50f01a28c9dd84a2410bb4b4
Public site status site-status.json	12 Jun 2026, 7:18 pm	4.3 KB	65ee441ba3884772fcfb3e7b53dd3bade5abd4ba58d183714969b9ff8a859621
Public release history releases.json	12 Jun 2026, 7:18 pm	10.0 KB	3aafd04f61047e656668fa747dceecb95a8ee02034a7afcd93bda20b2b446c41
Current cluster validation backtests/current-validation.json	12 Jun 2026, 5:53 pm	9.5 KB	fededec10bb9eb7ddd175dc76a2b62426b8438c6e9b5786f274b03ae2e1f8091
BLS crosswalk validation backtests/bls-crosswalk-validation.json	12 Jun 2026, 5:53 pm	3.3 KB	439c5168d76616f358dcf819e21c8b42e301e69b26760fb739d2cc9cf814646c
Multi-period temporal validation backtests/multi-period-validation.json	12 Jun 2026, 5:53 pm	5.5 KB	0bd080e777c63f7b4d1f29f56aa1c218f4cc52a4d4128450f916193d3c901204
Calibration diagnostics backtests/calibration-diagnostics.json	12 Jun 2026, 5:53 pm	5.6 KB	2b5bd77718394a320119bc31d6b2d4dc173b50ad87de56e1b5094e683066063b
Occupation-family validation backtests/occupation-family-validation.json	12 Jun 2026, 5:53 pm	5.6 KB	5ebb89d60047d098bd958ce6e1c586aab0619d663fbe3e999420dd6e176c19aa
Forecast-horizon sidecar backtests/forecast-horizon-validation.json	12 Jun 2026, 5:53 pm	3.3 KB	325c0c3f8047a93c3bb555482417ef7307eb5ceeeb4feb478fde6364c3cf2d61
IMF Singapore convergence backtests/imf-convergence.json	12 Jun 2026, 5:53 pm	3.7 KB	4d52dbf32f78585687517edbf5e9b61c22447689aaac3802686b139704c7bea4
Sensitivity analysis backtests/sensitivity-analysis.json	12 Jun 2026, 5:53 pm	18.5 KB	fa1710b340ed7c047a14b1c7bd31977d1544e67b8a1751d259cf8a5aafb10f6b
Occupation confidence ratings confidence-ratings.json	12 Jun 2026, 6:03 pm	531.2 KB	f308b45e1a8a1fc430589fb0e3a320328265c99f9b8fafe1424b8a4d0fb5a41b
Scenario families scenario-families.json	12 Jun 2026, 7:18 pm	752.6 KB	dabcd792ce7d6259205a425e5dba0bec275bfbdb6d6919ab385a8bd747e10aee
Singapore adoption diffusion adoption-diffusion.json	12 Jun 2026, 7:18 pm	2.6 KB	d4b08a91cca9d42eb51c9e92d49927c01d8f439a82036408772fec02752d9aca
Occupation age structure age-structure.json	12 Jun 2026, 7:18 pm	310.2 KB	7bed9487ec38339535ec29d956a1b21116f7c2f7dc8f380d03414ad555bf0b6e

Checksums are published so downloaded artifacts can be verified against the current release.

Major public claims are also published in a versioned claims matrix with evidence-strength labels, source keys, and research links. The citation layer is also published as research-library.json.

Dataset Citation

V7 public dataset

AI Work Index. V7 structural AI pressure dataset. 2026-06-11. https://aiworkindex.com/data

The structured Dataset metadata uses this same citation and links the current CSV, JSON, claims matrix, and field source map as machine-readable downloads.

Raw Data Health (32 valid, 0 issues)

Audit generated: 7 Jun 2026, 7:57 pm

Summary: 32 valid, 0 placeholder/error, 0 missing, 4 reference-only

Raw Input	Status	Used For	Notes
Postings source registry postings/source-registry.json	Valid	postings pipeline, ATS employer-source ingestion	Valid JSON payload.
Employer pressure seed registry employer-signals-seed.json	Valid	employer pressure pipeline	Valid JSON payload.
Singapore occupations base feed sg_occupations_complete_2024.json	Valid	score pipeline, occupation master list	Valid JSON payload.
Employment by occupation group employment_by_occupation.csv	Valid	score pipeline, market momentum	Header starts with: DataSeries,2025,2024,2023,2022,2021,2020,2019,2018,2017,2016,2015,2014,2013,2012,2011,2010,2009,2008,2007,2005,2004,2003
Median income by occupation group median_income_by_occupation.csv	Valid	score pipeline, market momentum, forecast-readiness wage outcome candidate	Header starts with: DataSeries,2023,2022,2021,2020,2019,2018,2017,2016,2015,2014,2013,2012,2011,2010,2009,2008,2007,2004,2003,2002,2001
MOM AI adoption among firms mom-ai-adoption-2026.json	Valid	AI in Singapore context, forecast-readiness adoption source	Valid JSON payload.
Vacancy rates by occupation group vacancy_rates_by_occupation_group.csv	Valid	labour monitor	Header starts with: quarter,industry1,industry2,occupation1,job_vacancy_rate
Job vacancies by industry and occupation job_vacancies_by_industry_and_occupation_quarterly.csv	Valid	industry context, labour monitor	Header starts with: DataSeries,20253Q,20252Q,20251Q,20244Q,20243Q,20242Q,20241Q,20234Q,20233Q,20232Q,20231Q,20224Q,20223Q,20222Q,20221Q,2021
Recruitment and resignation rates recruitment_resignation_rates.json	Valid	labour monitor hiring signal	Valid JSON payload.
Recruitment and resignation CSV recruitment_resignation_rates.csv	Reference only	labour monitor troubleshooting	Optional reference/backstop file is not present locally.
Retrenchment by occupation group retrenchment_by_occupation_group.json	Valid	labour monitor retrenchment signal	Valid JSON payload.
Retrenchment by occupation group CSV retrenchment_by_occupation_group.csv	Reference only	labour monitor troubleshooting	Optional reference/backstop file is not present locally.
Labour Force 2024 Section D LFR2024_SectionD.xlsx	Valid	worker profile, employment basis	Workbook sheets: Contents, D1, D2, D3, D4
Labour Force 2025 Section D LFR2025_SectionD.xlsx	Valid	worker profile, employment basis, delta validation	Workbook sheets: Contents, D1, D2, D3, D4
Industry x occupation employment industry_x_occupation.csv	Valid	industry context, industry momentum	Header starts with: DataSeries,2025,2024,2023,2022,2021,2020,2019,2018,2017,2016,2015,2014,2013,2012,2011,2010,2009,2008,2007,2006,2005,2004
Wages by industry wages_by_industry.xlsx	Valid	sector wage anchors	Workbook sheets: Contents, T4, T4.1, T4.2, T4.3
Wages by sex wages_by_sex.xlsx	Valid	worker profile	Workbook sheets: Contents, T1, T1.1, T1.2
Planning area of residence by occupation planning_area_residence_by_occupation.json	Valid	geography context	Valid JSON payload.
Travel time to work by occupation travel_time_to_work_by_occupation.json	Valid	geography context	Valid JSON payload.
Quarterly unemployment rate unemployment_rate_quarterly.json	Valid	macro context	Valid JSON payload.
Job vacancy to unemployed ratio job_vacancy_to_unemployed_ratio.json	Valid	macro context	Valid JSON payload.
WSQ trainees by statement attainment wsq_trainees_by_statement_attainment.json	Valid	transition infrastructure	Valid JSON payload.
Jobs Transformation Maps overview jtm_overview.html	Valid	transition infrastructure	Present locally.
Occupation list workbook occupations_list.xlsx	Reference only	reference	Reference/backstop file present locally, not currently used in the live pipeline.
Felten AIOE appendix AIOE_DataAppendix.xlsx	Valid	exposure ensemble	Workbook sheets: Index, Appendix A, Appendix B, Appendix C, Appendix D
Anthropic job exposure anthropic_job_exposure.csv	Valid	exposure ensemble	Header starts with: occ_code,title,observed_exposure
Eloundou occupation exposure eloundou_gpts_occ_level.csv	Valid	exposure ensemble	Header starts with: O*NET-SOC Code,Title,dv_rating_alpha,dv_rating_beta,dv_rating_gamma,human_rating_alpha,human_rating_beta,human_rating_ga
ILO refined exposure index ilo_genai_scores_isco08_2025.xlsx	Valid	exposure ensemble	Workbook sheets: Sheet1
Anthropic task penetration anthropic_task_penetration.csv	Valid	experimental task-primitives, V4.3 shadow model	Header starts with: task,penetration
O*NET occupation data raw/external/onet/Occupation_Data.txt	Valid	O*NET enrichment	Present locally.
O*NET task statements raw/external/onet/Task_Statements.txt	Valid	O*NET enrichment, experimental task-primitives	Present locally.
O*NET task ratings raw/external/onet/Task_Ratings.txt	Valid	experimental task-primitives, V4.3 shadow model	Present locally.
O*NET technology skills raw/external/onet/Technology_Skills.txt	Valid	O*NET enrichment	Present locally.
Singapore empirical mobility transitions sg_empirical_mobility.json	Valid	experimental transition model, V4.3 shadow model	Valid JSON payload.
BLS occupation projections bls_projections_2024_2034.xlsx	Valid	proxy employment, convergent check	Workbook sheets: Index, Table 1.1, Table 1.2, Table 1.3, Table 1.4
BLS CPS employment bls_cps_employment_2025.xlsx	Reference only	reference	Reference/backstop file present locally, not currently used in the live pipeline.

This audit distinguishes real local raw inputs from missing files and failed download artifacts.

Evidence Tiers

Tier	Meaning
Official local	Direct local government or official-statistics data published at the level shown on the site.
Derived from official local	Rule-based or aggregated fields anchored to official local data or published policy scope.
Cross-country research	Peer-reviewed or multi-country research inputs used for the structural baseline or calibration.
External proxy	Non-local research or external data used as an exposure input, proxy, or cross-check.
Synthetic	Estimated role constructs or illustrative outputs that do not map one-to-one to official occupations.

The employment estimate and wage-pool proxy are intentionally separated. The first is an estimated local allocation from official Labour Force 2025 2-digit occupation-family totals, weighted within each family; the second is a BLS-weighted proxy used only for wage-pool views. Separate live worker-profile context comes from Labour Force 2025 Section D and wages-by-sex tables.

Current labour evidence is also published separately in the reference-market context pack so the structural score and the monitor can be audited independently.

A note on data quality

All scores are deterministic and reproducible. The scoring pipeline uses no LLM in the loop, and every displayed proxy or estimate now carries an explicit basis and evidence tier in the dataset.

Source Registry (57 sources)

Live sources are tracked separately from the structural score. Some sources are already active in the live monitor pipeline; others remain reference or backlog sources and are not yet part of the published monitor.

Source	Tier	Status	Used For
MOM Occupational Wage Survey 2024 2024	Official local	Live	median wages, wage quartiles
MOM / SingStat Detailed SSOC Employment Counts not publicly released	Official local	Requested	official detailed occupation employment
MOM Labour Force 2024 Table D8 2024	Official local	Live	sub-major employment totals, estimated occupation employment basis
MOM Labour Force 2025 Table D8 2025	Official local	Live	sub-major employment totals, estimated occupation employment basis, family delta validation
MOM Employment by Occupation Group 2024	Official local	Live	major-group market momentum
MOM Industry × Occupation 2024	Official local	Live	industry momentum spread, Singapore context support, industry footprint
MOM Labour Force 2024 Section D 2024	Official local	Live	worker profile context, detailed gender anchors, estimated employment basis
MOM Labour Force 2025 Section D 2025	Official local	Live	worker profile context, detailed gender anchors, family delta context, industry workforce context, industry occupation mix context, experimental demand fragility overlay
SingStat Planning Area of Residence by Occupation 2020	Official local	Live	geography context, planning-area concentration
SingStat Travel Time to Work by Occupation 2020	Official local	Live	geography context, commute burden context
US BLS Occupational Projections 2024-2034 2024-2034	Official local	Live	US demand resilience, US employment outlook, US wage context
US BLS CPS Employment 2025 2025	Official local	Live	US current employment context, US calibration
US BLS OEWS 2025	Official local	Available	US wage context, US wage distribution, occupational pay calibration
US BLS Occupational Requirements Survey 2025	Official local	Available	US transition capacity, US work-context friction, US bottleneck calibration
US BLS CPS Demographics by Occupation 2025	Official local	Available	US worker profile, US demographic context, equity and transition framing
US BLS Skills Data 2025	Official local	Available	US skills context, US task and skill explanation
US Occupational Outlook Handbook 2025	Official local	Available	US narrative context, career guidance, work-environment copy
MOM Jobs in Demand 2025 2025	Official local	Live	demand signal, market resilience bonus
MOM Shortage Occupation List 2026 2026	Official local	Live	demand signal, market resilience bonus
MOM Labour Market Reports / monitor series 2025	Official local	Live	labour monitor, quarterly context, cluster vacancy rates
MOM Labour Market Report Q4 2025 2025 Q4	Official local	Live	labour monitor enrichment, vacancy validation, recruitment and resignation context, retrenchment incidence context, forecast-readiness gating
Job Vacancies by Industry and Occupation 2025	Official local	Live	industry context vacancy overlay, cluster vacancy counts
MOM Job Vacancy Rates by Occupation Group 2025	Official local	Live	labour monitor, vacancy validation
MOM Job Vacancy Counts 2025	Official local	Live	labour monitor, vacancy validation
MOM Recruitment and Resignation Rates 2025	Official local	Live	labour monitor hiring signal
MOM / SingStat Retrenchment by Occupation Group 2026	Official local	Live	labour monitor retrenchment signal
Median Gross Monthly Income by Occupation and Sex 2023	Official local	Live	wage movement source candidate, raw data audit
Singapore postings monitor 2026	External proxy	Live	postings volume proxy, AI-skill share candidate, occupation and role context
SingStat Unemployment Rate, Quarterly, Seasonally Adjusted 2025	Official local	Live	macro context, unemployment context
MOM Job Vacancy to Unemployed Person Ratio 2024	Official local	Live	macro context, labour tightness context
Workforce Singapore Jobs Transformation Maps 2026	Official local	Live	transition infrastructure, transition-support context
SkillsFuture WSQ Trainees by Statement Attainment 2024	Official local	Live	transition infrastructure
SkillsFuture Career Transition Programme 2026	Official local	Live	transition infrastructure, transition-support context
Workforce Singapore Career Conversion Programmes 2026	Official local	Live	transition infrastructure, transition-support context
Workforce Singapore CareersFinder 2025	Official local	Live	transition infrastructure, transition-support context
Wages by Industry 2024	Official local	Live	sector wage anchors, industry context support
Wages by Sex 2024	Official local	Live	worker-profile wage context
IMDA Singapore Digital Economy Report 2025 2025	Official local	Live	AI in Singapore context, reports, national AI adoption framing
MOM AI Adoption Among Firms 2026	Official local	Live	AI in Singapore context, near-term adoption context, forecast-readiness gating
IMDA National AI Impact Programme 2026	Official local	Live	AI in Singapore context, national programme framing
MOM Singapore Opportunity Index 2025	Official local	Live	AI in Singapore context, employer-quality context, career mobility framing
Felten AIOE 2021	External proxy	Live	exposure ensemble
Pizzinelli et al. theta 2023	External proxy	Live	human bottleneck
Anthropic Economic Index 2026-01	External proxy	Live	exposure ensemble, observed usage gap
Eloundou GPT Exposure 2023	External proxy	Live	exposure ensemble
ILO Generative AI Exposure Index 2025	External proxy	Live	exposure ensemble
O*NET Occupation Data 2026	External proxy	Live	task and technology enrichment context
O*NET Task Statements 2026	External proxy	Live	task primitive matching, task enrichment context
O*NET Task Ratings 2026	External proxy	Live	task primitive weighting
Anthropic task penetration 2026	External proxy	Live	task primitive penetration, V7 task signal
O*NET Technology Skills 2026	External proxy	Live	task and technology enrichment context
O*NET Job Zones 2026	External proxy	Live	education label proxy
US BLS Occupational Projections 2024-2034 2024-2034	External proxy	Live	employment proxy, convergent cross-check
Local wage-floor mapping 2026	Derived from official local	Live	local context
Local licensing mapping 2026	Derived from official local	Live	local context
Labour-dependency mapping 2026	Derived from official local	Live	local context
Transition-program mapping 2026	Derived from official local	Live	local context

Data Dictionary

69 fields across 9 categories. See the methodology page for derivation details.

Identity & classification (4 fields)

Field	Type	Description
ssoc	string	[Singapore] Singapore Standard Occupational Classification code (5-digit detail code).
title	string	Occupation title from the national classification (MOM for Singapore).
major_group	string	Major occupational group key (for example professionals or managers).
education_label	string	Displayed education proxy. Derived from O*NET Job Zones via the country crosswalk.

Wages & employment (11 fields)

Field	Type	Description
gross_wage_median	number	Median gross monthly wage in local currency from MOM 2024 (Singapore).
gross_wage_25th	number	25th percentile gross monthly wage in local currency.
gross_wage_75th	number	75th percentile gross monthly wage in local currency.
employment_thousands	number	Legacy compatibility alias for estimated_sg_employment_thousands. Derived from published Labour Force 2025 2-digit occupation-family totals and weighted within each family. Not an official occupation headcount.
estimated_sg_employment_thousands	number	[Singapore only] Est. Singapore employment for this occupation, derived from published Labour Force 2025 2-digit occupation-family totals and weighted within each family using BLS proxy employment plus Singapore wage information. Not an official occupation headcount.
employment_basis	enum	Basis label for estimated_sg_employment_thousands. Current live basis: Est. local sub-major weighted allocation (2025).
bls_proxy_employment	number	BLS-weighted proxy employment used for wage-pool analysis. This is not an official Singapore occupation headcount.
data_basis.employment_estimate	object	Tier/source metadata for the Singapore employment estimate.
data_basis.wage_pool_proxy	object	Tier/source metadata for the wage-pool proxy field.
labour_monitor_key	enum\|null	Cluster key linking this occupation to the separately published Singapore labour monitor dataset.
group_employment_thousands	number	Official employment total for the broad Singapore occupation group (in thousands).

Core scoring (13 fields)

Field	Type	Description
exposure	number	Base 4-source exposure ensemble (0-1), before the V7 task-concentration buffer.
task_signal	number	V7 task-concentration signal: task_effective_coverage × task_exposure_concentration. Zero when task primitives are unavailable.
exposure_v7	number	Live V7 exposure after the task-concentration buffer (Hampole et al. 2025: concentration offsets labour-demand losses). Formula: min(1, exposure × (1 − 0.20 × task_signal)).
bottleneck	number	Human bottleneck strength (0-1). Higher means stronger human advantage from judgment, accountability, or interpersonal work.
displacement_pressure	number	Intermediate structural pressure field (0-1). Formula: exposure_v7 × (1 - bottleneck).
demand_signal_bonus	number	Additive demand bonus from exact or prefix matches against SOL and Jobs in Demand.
demand_persistence	number	V7 demand-side counterforce proxy built from market momentum, vacancy rank, scarcity rank, and official demand-signal rank.
demand_resilience	number	Published V7 demand-resilience field (0-1). Formula: min(1, base_resilience × 0.45 + demand_signal_bonus + 0.10 × demand_persistence).
net_risk	number	Headline displacement risk (0-1). Formula: headline_risk = displacement_pressure × (1 - demand_resilience), where displacement_pressure = exposure_v7 × (1 - bottleneck).
risk_band	enum	Categorical risk: very_low (<5%), low (<15%), moderate (<30%), high (<50%), very_high (≥50%).
augmentation	number	Live V7 augmentation potential (0-1). Formula: exposure_v7 × bottleneck × market.market_resilience.
impact_type	enum	ai_leveraged \| at_risk \| stable \| mixed, based on displacement and augmentation thresholds.
baseline_v6	object	Retained V6 baseline fields for release-to-release comparison. Contains baseline_v6.net_risk and baseline_v6.exposure.

Market signals (5 fields)

Field	Type	Description
market.market_momentum	number	Broad labour-market momentum signal (0-1).
market.industry_footprint_momentum	number\|null	Industry-footprint employment momentum when occupation-specific industry context is available.
market.market_resolution	enum	How market resilience was resolved for the occupation, for example broad-only or industry-footprint blend.
market.occupation_scarcity	number	Labour shortage signal (0-1), derived from SOL and Jobs in Demand lists.
market.market_resilience	number	Combined market buffer (0-1). Higher means stronger demand protection.

Evidence & provenance (12 fields)

Field	Type	Description
evidence.anthropic_calibrated	boolean	Whether Anthropic observed-usage data is available for this occupation.
evidence.anthropic_gap	number\|null	Observed-usage percentile minus theoretical exposure percentile.
evidence.anthropic_observed_pctile	number\|null	Observed Anthropic usage percentile when a direct Anthropic match exists.
evidence.sol_match	string\|false	Shortage Occupation List match: exact, prefix, or false.
evidence.jobs_in_demand_match	string\|false	Jobs in Demand match: exact, prefix, or false.
evidence.exposure_blend_strategy	enum	Current exposure provenance marker for the deterministic audited 4-source stack.
evidence.exposure_agreement	enum	Agreement across matched exposure sources: consensus_high, consensus_low, aligned_mid, divergent, or insufficient_data.
evidence.exposure_source_weights	object	Normalized per-source weights used inside the exposure ensemble after conditioning on which sources matched this occupation.
evidence.exposure_source_keys	string[]	List of matched exposure sources used by the audited ensemble.
evidence.exposure_source_pctiles	object	Persisted per-source exposure percentiles for the matched AIOE, Anthropic, Eloundou, and ILO inputs.
evidence.signal_conflict	boolean	Whether the occupation has materially conflicting evidence, such as high structural pressure but strong current demand signals.
evidence.signal_conflict_reasons	string[]	Explicit reasons why the occupation is flagged as having conflicting evidence.

Confidence & uncertainty (14 fields)

Field	Type	Description
confidence.score	number	Overall estimate confidence (0-1), combining crosswalk quality, market data, and source freshness.
confidence.exposure_source_count	number	How many exposure sources were available in the audited 4-source exposure stack.
confidence.source_coverage	number	Coverage component of confidence (0-1), based on the number of matched exposure sources.
confidence.signal_agreement	number	Agreement component of confidence (0-1), derived from agreement across available exposure sources.
confidence.sensitivity	number	Sensitivity component of confidence (0-1), derived from the stability stress test.
confidence.level	enum	Final published confidence label. This may be capped below the raw threshold result for fallback, sparse-source, or contested cases.
confidence.threshold_level	enum	Raw threshold label implied by confidence.score before any publication policy cap is applied.
confidence.policy_cap_reason	enum\|null	If present, explains why the published label is more conservative than the raw threshold: insufficient_source_count, fallback_mapping, major_fallback_mapping, or signal_conflict.
uncertainty.exposure_p10	number	10th-percentile interval estimate for exposure.
uncertainty.exposure_p50	number	Median interval estimate for exposure.
uncertainty.exposure_p90	number	90th-percentile interval estimate for exposure.
uncertainty.net_risk_p10	number	10th-percentile interval estimate for net_risk.
uncertainty.net_risk_p50	number	Median interval estimate for net_risk.
uncertainty.net_risk_p90	number	90th-percentile interval estimate for net_risk.

Task primitives (4 fields)

Field	Type	Description
task_primitives.matched_task_weight_share	number\|null	Share of weighted O*NET task importance/frequency that matched the experimental Anthropic task-penetration layer.
task_primitives.task_effective_coverage	number\|null	Experimental task-weighted effective coverage sidecar. Null when weighted task portfolios are not available.
task_primitives.task_exposure_concentration	number\|null	Experimental task-weighted concentration sidecar. Higher means exposure is concentrated into fewer important tasks.
task_primitives.method	enum\|null	Experimental task-primitive method identifier. Null when no weighted task evidence is available for this occupation.

Context & stability (2 fields)

Field	Type	Description
sg_context	object	Local context flags (policy coverage, licensing, labour dependency, transition support).
stability.label	enum	stable \| watch \| sensitive — how much the risk band moves under the Monte Carlo stability check.

Workflow overlay (1 fields)

Field	Type	Description
workflow_overlay.*	object	Heuristic workflow dimensions used by role and scenario tooling. These provide interpretive context rather than direct live-score inputs.