Datasets:

Cespial
/

adi

isco08 large_string	titulo large_string	adi_l2 float64	adi_l3 float64	adi_share float64	n_tareas int64	barrera_fisica float64	barrera_responsabilidad float64	barrera_confianza float64	barrera_acceso float64	ic_lo float64	ic_hi float64	sintetico bool
1111	Legislators	0.5	0	42.17	30	0.2667	0.5333	0.8	0.5167	36.17	48.25	false
1112	Senior government officials	0.7612	0	50.68	89	0.2081	0.4927	0.6208	0.6613	46.79	54.51	false
1113	Traditional chiefs and heads of village	0.6276	0	47.01	79	0.1396	0.4819	0.7056	0.5741	43.28	50.47	false
1114	Senior officials of special-interest organisations	0.9033	0.0097	55.17	305	0.1977	0.3457	0.4805	0.741	53.54	57.01	false
1120	Managing directors and chief executives	0.7586	0	50.34	66	0.0632	0.4079	0.6351	0.6486	46.85	54.08	false
1211	Finance managers	0.8676	0.0288	56.48	39	0.0476	0.6007	0.3376	0.761	52.02	60.65	false
1212	Human resource managers	0.8953	0	56.63	60	0.0206	0.3203	0.6688	0.7567	53.83	59.36	false
1213	Policy and planning managers	0.886	0.0151	55.17	252	0.2416	0.4882	0.4402	0.7895	53.4	56.75	false
1219	Business services and administration managers not elsewhere classified	0.8065	0.0122	53.01	371	0.2714	0.3902	0.4751	0.7221	50.53	55.69	false
1221	Sales and marketing managers	0.8768	0.031	53.87	37	0.0966	0.0776	0.5201	0.7239	49.79	58.04	false
1222	Advertising and public relations managers	0.8987	0.0162	55.57	66	0.0777	0.0487	0.6005	0.5757	51.83	59.01	false
1223	Research and development managers	0.8725	0	52.13	107	0.2542	0.4468	0.4129	0.7405	49.87	54.23	false
1311	Agricultural and forestry production managers	0.5654	0.0374	44.05	30	0.7507	0.2858	0.1922	0.8429	38.47	50.25	false
1312	Aquaculture and fisheries production managers	0.5654	0.0374	44.05	30	0.7507	0.2858	0.1922	0.8429	38.36	50.59	false
1321	Manufacturing managers	0.6963	0.0114	47.76	116	0.5083	0.6033	0.3158	0.9045	44.97	50.62	false
1322	Mining managers	0.886	0.0151	55.17	252	0.2416	0.4882	0.4402	0.7895	53.39	56.87	false
1323	Construction managers	0.7793	0	48.81	25	0.516	0.6735	0.5204	0.7697	43.21	54.2	false
1324	Supply, distribution and related managers	0.9011	0	55.21	30	0.2756	0.2787	0.4391	0.8729	50.48	59.81	false
1330	Information and communications technology service managers	0.9342	0	54.07	17	0	0.2013	0.4196	0.605	47.8	59.56	false
1341	Child care services managers	0.7001	0	47.65	17	0.3043	0.5872	0.627	0.6442	38.14	56.62	false
1342	Health services managers	0.7137	0.0596	52.21	18	0.1417	0.4733	0.5087	0.7427	45.07	59.19	false
1343	Aged care services managers	0.7602	0.0464	51.49	35	0.1438	0.422	0.5371	0.7544	45.96	57.47	false
1344	Social welfare managers	0.68	0	48.53	16	0.0687	0.2864	0.8315	0.6073	43	55.04	false
1345	Education managers	0.7478	0	48.89	83	0.2386	0.4509	0.6513	0.7152	45.39	52.27	false
1346	Financial and insurance services branch managers	0.8745	0.0252	55.54	56	0.0605	0.5559	0.375	0.7652	51.37	59.4	false
1349	Professional services managers not elsewhere classified	0.886	0.0151	55.17	252	0.2416	0.4882	0.4402	0.7895	53.46	56.86	false
1411	Hotel managers	0.8107	0	50.44	24	0.5545	0.0802	0.6042	0.6863	43.73	56.33	false
1412	Restaurant managers	0.6921	0.0662	45.92	28	0.5728	0.1988	0.3058	0.6738	36.23	55.73	false
1420	Retail and wholesale trade managers	0.9231	0	48.96	17	0.1511	0.2421	0.6368	0.7951	39.87	56.16	false
1431	Sports, recreation and cultural centre managers	0.8652	0.009	54.66	271	0.2893	0.5605	0.4223	0.8046	51.65	57.63	false
1439	Services managers not elsewhere classified	0.886	0.0151	55.17	252	0.2416	0.4882	0.4402	0.7895	53.49	56.85	false
2111	Physicists and astronomers	0.8078	0.0676	56.32	51	0.1777	0.1925	0.3045	0.4151	50.34	61.98	false
2112	Meteorologists	0.9437	0.0743	60.35	51	0.1269	0.2237	0.1666	0.4007	56.27	64.72	false
2113	Chemists	0.6759	0	46.73	52	0.6028	0.3322	0.1613	0.8332	41.87	52.45	false
2114	Geologists and geophysicists	0.8266	0	52.91	81	0.3984	0.5796	0.1504	0.6834	49.53	56.08	false
2120	Mathematicians, actuaries and statisticians	0.9646	0.0528	61.73	125	0.0187	0.2652	0.2508	0.5154	59.62	63.91	false
2131	Biologists, botanists, zoologists and related professionals	0.7197	0.0203	47.49	219	0.5737	0.3438	0.2721	0.7393	45.26	49.58	false
2132	Farming, forestry and fisheries advisers	0.6844	0	46.51	67	0.6943	0.3187	0.4995	0.7191	42.71	50.8	false
2133	Environmental protection professionals	0.8168	0.0132	53.19	163	0.3938	0.3864	0.3533	0.6663	50.6	55.76	false
2141	Industrial and production engineers	0.8764	0	54.03	46	0.3376	0.1388	0.3233	0.879	49.67	58.01	false
2142	Civil engineers	0.7853	0	50.92	42	0.4132	0.9347	0.1937	0.7891	46.09	55.58	false
2143	Environmental engineers	0.8856	0.0102	55.33	57	0.2394	0.7906	0.2799	0.7975	52.26	58.37	false
2144	Mechanical engineers	0.8492	0.0145	51.24	139	0.3998	0.5405	0.2121	0.828	48.91	53.52	false
2145	Chemical engineers	0.8888	0	52.7	14	0.5721	0.555	0.0982	0.9697	46.31	58.37	false
2146	Mining engineers, metallurgists and related professionals	0.7862	0	49.17	78	0.4552	0.6291	0.179	0.9057	45.8	52.03	false
2149	Engineering professionals not elsewhere classified	0.8936	0.0132	53.24	336	0.4667	0.498	0.1762	0.8279	51.83	54.64	false
2151	Electrical engineers	0.9176	0	54.22	22	0.3894	0.6051	0.2186	0.7224	48.98	59.08	false
2152	Electronics engineers	0.8555	0.0408	53.53	59	0.3267	0.1415	0.1781	0.7858	49.88	56.87	false
2153	Telecommunications engineers	0.9006	0.0174	55.48	41	0.3838	0.283	0.176	0.8198	51.84	59.33	false
2161	Building architects	0.9114	0	51.83	24	0.182	0.7864	0.366	0.5893	47.99	55.4	false
2162	Landscape architects	0.8037	0	51.71	19	0.3348	0.4917	0.4151	0.5771	45.92	57.68	false
2163	Product and garment designers	0.7581	0	48.09	37	0.429	0.0209	0.3392	0.4193	43.72	52.72	false
2164	Town and traffic planners	0.8224	0	56.21	25	0.1263	0.4126	0.47	0.6691	50.49	61.52	false
2165	Cartographers and surveyors	0.8366	0	56.79	54	0.2827	0.6768	0.0963	0.6823	53.27	60.15	false
2166	Graphic and multimedia designers	0.9679	0	60.41	32	0.0793	0	0.0255	0.1157	57.37	63.71	false
2211	Generalist medical practitioners	0.5622	0.0021	41.24	300	0.5043	0.9513	0.7295	0.856	39.26	43.15	false
2212	Specialist medical practitioners	0.5292	0.0017	38.33	330	0.5014	0.9558	0.6864	0.8644	36.26	40.62	false
2221	Nursing professionals	0.4442	0	35.89	188	0.5681	0.8912	0.6898	0.8543	33.48	38.23	false
2222	Midwifery professionals	0.337	0	36.11	21	0.5723	0.892	0.761	0.6658	26.99	44.5	false
2230	Traditional and complementary medicine professionals	0.4784	0.0099	35.28	54	0.5696	0.906	0.7935	0.5827	29.33	40.88	false
2240	Paramedical practitioners	0.3559	0	28.19	28	0.7857	0.9522	0.594	0.8286	20.95	35.01	false
2250	Veterinarians	0.2841	0.0385	31.03	21	0.7456	0.8673	0.5685	0.4679	22.3	41.29	false
2261	Dentists	0.2718	0	24.76	57	0.7829	0.9548	0.7646	0.6036	19.56	30.04	false
2262	Pharmacists	0.7757	0	49.17	20	0.4583	0.8793	0.4123	0.8634	41.27	56.14	false
2263	Environmental and occupational health and hygiene professionals	0.6951	0	50.27	38	0.3868	0.4907	0.4717	0.8076	44.7	55.62	false
2264	Physiotherapists	0.4672	0	39.26	49	0.6037	0.7487	0.8087	0.5079	33.11	45.44	false
2265	Dieticians and nutritionists	0.8703	0	54.15	28	0.2439	0.7662	0.5671	0.7109	49.66	58.53	false
2266	Audiologists and speech therapists	0.4701	0	40.7	45	0.4849	0.8747	0.786	0.6602	35.92	45.63	false
2267	Optometrists and ophthalmic opticians	0.3072	0	34	10	0.638	0.9382	0.7401	0.5483	24.9	43.47	false
2269	Health professionals not elsewhere classified	0.4595	0.0011	37.5	199	0.541	0.8432	0.8208	0.6153	34.46	40.47	false
2310	University and higher education teachers	0.6804	0.0073	49.55	860	0.2636	0.382	0.5949	0.5902	48.22	50.74	false
2320	Vocational education teachers	0.5349	0	42.43	84	0.5822	0.4079	0.6592	0.6053	38.23	46.64	false
2330	Secondary education teachers	0.4987	0	41.58	32	0.5286	0.3791	0.6083	0.6818	35.2	48.15	false
2341	Primary school teachers	0.4301	0	37.83	73	0.6547	0.3373	0.7126	0.5942	33.9	42.44	false
2342	Early childhood educators	0.1813	0	26.58	71	0.8412	0.3484	0.796	0.4442	23.11	30.43	false
2351	Education methods specialists	0.8552	0	58.4	30	0.0909	0.1445	0.419	0.4687	53.86	62.32	false
2352	Special needs teachers	0.332	0	34.43	196	0.6804	0.5262	0.7742	0.6072	31.89	36.94	false
2353	Other language teachers	0.5702	0.0208	44.75	104	0.4361	0.2186	0.6817	0.5225	40	48.92	false
2354	Other music teachers	0.5921	0.0466	46.6	30	0.4469	0.1361	0.6796	0.4434	39.81	53.53	false
2355	Other arts teachers	0.4816	0.0457	40.7	67	0.5111	0.2631	0.7065	0.5278	34.1	47.26	false
2356	Information technology trainers	1	0	61.45	20	0.0557	0.0199	0.4087	0.688	57.47	65.26	false
2359	Teaching professionals not elsewhere classified	0.5505	0.026	43.04	72	0.3757	0.3639	0.7743	0.5853	37.74	48.49	false
2411	Accountants	0.9699	0.093	66.59	54	0.0277	0.5244	0.2292	0.6366	63.32	69.64	false
2412	Financial and investment advisers	0.9523	0.0231	61.77	77	0.0112	0.6797	0.6499	0.6398	58.9	64.38	false
2413	Financial analysts	0.939	0.0844	61.82	84	0.0317	0.6245	0.26	0.6968	56.56	66.27	false
2421	Management and organisation analysts	0.9253	0.0347	61	94	0.1311	0.0693	0.3235	0.7834	58.47	63.52	false
2422	Policy administration professionals	0.906	0.0902	61.87	156	0.215	0.3024	0.1811	0.7501	59.41	64.32	false
2423	Personnel and careers professionals	0.8486	0.0189	57.14	111	0.0424	0.3064	0.5374	0.6888	54.05	59.88	false
2424	Training and staff development professionals	1	0	61.45	20	0.0557	0.0199	0.4087	0.688	57.33	65.1	false
2431	Advertising and marketing professionals	0.9723	0.0699	67.14	49	0.0068	0.0443	0.206	0.4817	63.87	70.12	false
2432	Public relations professionals	0.9298	0	58.09	18	0.0173	0.0593	0.6242	0.4814	51.62	63.99	false
2433	Technical and medical sales professionals (excluding ICT)	0.895	0.0628	59.4	64	0.1669	0.1384	0.5275	0.5862	55.49	62.91	false
2434	Information and communications technology sales professionals	0.9082	0.0363	57.81	89	0.1826	0.1277	0.5667	0.6902	54.47	61.25	false
2511	Systems analysts	0.943	0	56.43	54	0.0413	0.1268	0.4183	0.6852	53.75	58.87	false
2512	Software developers	0.9698	0.1278	64.31	47	0.0133	0	0.1897	0.4858	60	69.04	false
2513	Web and multimedia developers	0.9511	0.3467	70.97	59	0	0	0.1034	0.282	67.05	74.48	false
2514	Applications programmers	0.9697	0.3381	68.64	17	0.0234	0	0.2285	0.2363	60.7	74.94	false
2519	Software and applications developers and analysts not elsewhere classified	0.983	0.0569	63.77	415	0.0385	0.0957	0.1531	0.6518	62.78	64.84	false
2521	Database designers and administrators	1	0.0638	64.57	43	0	0.0404	0.0728	0.7205	62.32	66.95	false
2522	Systems administrators	1	0.0459	60.32	20	0.3813	0.0484	0.0777	0.8894	56.97	63.78	false

End of preview.

Agent Delegability Index (ADI) v2026.1

35.2 % of task time in world employment is at least partly delegable to AI agents (L2 or L3) in 2026; only 2.5 % can be handed over in full (L3). Delegable employment rises with income — 22.5 % in low-income, 28.3 % in lower-middle, 35.0 % in upper-middle and 48.7 % in high-income countries — the delegation divide (7.23 percentage points per log point of GDP per capita, 146 countries).

Dataset of the paper The Delegation Divide: An Open Index of AI-Agent Delegability across Occupations and Countries (Espinal Maya & Jiménez Londoño, 2026). Preregistered at OSF: doi:10.17605/OSF.IO/9ZNTF. Explore every occupation and country, with charts and citations: https://balamlabs.org/adi (Balam Labs). Dataset DOI: 10.5281/zenodo.23552918. Code, rubric, MCP server and full pipeline: https://github.com/Cespial/delegation-divide.

Configs

config	rows	unit
`adi_isco` (default)	433	ISCO-08 unit group (4 digits)
`adi_pais`	146	country (ISO 3166-1 alpha-3)
`adi_tarea`	18,796	O*NET 30.2 task statement, with each model's level

from datasets import load_dataset
occ = load_dataset("cespial/adi", "adi_isco", split="train")
countries = load_dataset("cespial/adi", "adi_pais", split="train")
tasks = load_dataset("cespial/adi", "adi_tarea", split="train")

The repository also carries the same tables as CSV, claims.jsonl (every number in the paper), the raw model outputs (ratings_sonnet.jsonl.gz, ratings_haiku.jsonl.gz), a static JSON API (api/isco/<code>.json, api/country/<ISO3>.json, api/index.json), DATA_DICTIONARY.md and SHA256SUMS.

Rating scale (verbatim from the frozen rubric)

L0 — No meaningful role. The agent cannot perform or materially speed up any substantial part of the task.
L1 — Augmentation. A human performs the task; the AI assists step by step (drafts, searches, suggestions). The human stays in the loop at each step and produces the output.
L2 — Partial delegation. An agent can execute substantial sub-steps of the task end to end from a goal specification. A human still has to perform at least one essential step or review and correct the work before use.
L3 — Full delegation. An agent can complete the whole task end to end, to the quality a competent worker would accept, given only the goal and access. The human's role is limited to assigning it and accepting the result.

Reference technology: publicly available AI agents as of October 2026 (tool use, code execution, browsing, computer use, file handling; no physical body). Ratings measure technical feasibility, not adoption or legality. Each task also carries the share of working time an agent could take over, the minutes a human needs, and four barrier flags (physical, accountability, trust, access).

Main columns

adi_isco: isco08, titulo, adi_l2 (share 0–1 of task weight at L2+), adi_l3 (share at L3), adi_share (% of working time delegable), n_tareas, barrera_{fisica,responsabilidad,confianza,acceso}, ic_lo/ic_hi (95 % bootstrap interval of adi_share).
adi_pais: iso3, pais, ingreso (World Bank FY2026), anio, resolucion (ISCO level of the employment data), adi_l2, adi_l3, adi_share (employment-weighted), empleo_total.
adi_tarea: task_id (O*NET), titulo, texto, consensus nivel, share, minutos, barrier fractions, nivel_sonnet, nivel_haiku.

Types, units and definitions for every column: DATA_DICTIONARY.md.

How it was built

Raters: Claude Sonnet 5.5 (primary) and Claude Haiku 5.5 (secondary), one task per call, neutral configuration, rubric hash checked before every run. Consensus: lower level when the models disagree (quadratic-weighted κ = 0.77 between them). Tasks are weighted by O*NET importance × frequency and carried to ISCO-08 through O*NET-SOC 2019 → SOC 2010 → ISCO-08 (433 of 436 unit groups). Country scores weight unit groups by ILOSTAT employment (latest year ≥ 2019, 2-digit or 1-digit ISCO) with within-group four-digit weights from BLS OES (high income) or Colombia's GEIH (other groups).

Limitations

Ratings come from two models of one developer; a blind human rating of 300 tasks agrees with the model consensus at kappa = 0.85, and an agentic panel on 1,500 tasks at kappa = 0.87 (see the paper, section 7).
O*NET describes work in the United States; the same task mix is carried to every country, so cross-country differences are compositional.
Technical delegability is an upper bound on delegation, not a forecast of employment effects.

Licence and attribution

CC BY 4.0. Includes information from the O*NET 30.2 Database by the U.S. Department of Labor, Employment and Training Administration (USDOL/ETA), used under the CC BY 4.0 license (task statements are reproduced verbatim in adi_tarea). ESCO, ILOSTAT, World Bank, BLS and DANE inputs are used only as derived aggregates.

Citation

@article{espinal2026delegation,
  title   = {The Delegation Divide: An Open Index of AI-Agent Delegability across Occupations and Countries},
  author  = {Espinal Maya, Cristian and Jim{\'e}nez Londo{\~n}o, Santiago},
  journal = {arXiv preprint},
  year    = {2026},
  url     = {https://github.com/Cespial/delegation-divide},
  note    = {Preregistration: doi:10.17605/OSF.IO/9ZNTF}
}

Downloads last month: 10