← Back to the map

How many jobs will AI replace by 2030?

A transparent, reproducible scenario estimate across 46 countries — not a single prediction. We weight each occupation’s generative-AI exposure (ILO Working Paper 140) by its automation share and three realisation rates calibrated to published research.

Global total across 46 countries · ~2391M workers covered. Mid scenario is calibrated to ~3% of the workforce, in line with Goldman Sachs and WEF estimates (below).

AI job loss by 2030, by country (mid scenario)

Country	Workforce	Est. jobs lost (mid)	Low–High
China	750,642,741	21,765,772	10,882,884–47,884,689
India	551,859,980	13,150,796	6,575,403–28,931,762
Indonesia	170,035,856	6,164,065	3,082,015–13,560,930
United States	146,791,230	4,955,770	2,477,912–10,902,709
Japan	99,222,400	3,205,904	1,602,954–7,053,007
Brazil	102,196,347	3,097,731	1,548,860–6,814,997
Germany	55,164,100	1,977,280	988,638–4,350,008
Vietnam	56,173,070	1,960,142	980,061–4,312,282
Mexico	56,702,659	1,667,137	833,566–3,667,694
South Korea	33,346,523	1,185,077	592,542–2,607,172
France	36,409,250	1,109,446	554,732–2,440,802
United Kingdom	30,865,300	1,090,811	545,404–2,399,779
Thailand	37,103,969	974,767	487,366–2,144,473
Turkey	30,650,334	923,638	461,819–2,032,029
Italy	28,475,450	865,233	432,626–1,903,506
Spain	26,602,850	817,840	408,910–1,799,252
Canada	21,540,980	770,754	385,371–1,695,647
Argentina	20,819,935	755,698	377,845–1,662,542
Malaysia	17,792,300	650,486	325,239–1,431,065
Poland	15,879,607	530,158	265,068–1,166,344
Australia	14,562,050	519,609	259,796–1,143,124
Netherlands	10,735,800	358,134	179,069–787,895
Chile	9,414,852	334,465	167,235–735,807
Romania	7,234,871	212,619	106,296–467,781
Sweden	5,200,004	187,955	93,973–413,499
Switzerland	4,604,736	172,911	86,460–380,431
Belgium	4,770,471	172,753	86,371–380,054
Portugal	4,924,235	168,425	84,215–370,529
Czechia	4,898,773	161,319	80,666–354,909
Austria	4,195,998	147,004	73,511–323,410
Hungary	4,369,244	142,299	71,154–313,067
Greece	4,024,989	141,312	70,662–310,917
Ireland	3,130,140	104,547	52,271–229,994
New Zealand	3,058,070	102,515	51,247–225,539
Norway	2,710,864	87,181	43,592–191,810
Slovakia	2,427,232	83,749	41,873–184,245
Singapore	2,114,426	80,743	40,364–177,617
Finland	2,313,806	73,896	36,945–162,585
Denmark	1,688,763	58,804	29,405–129,395
Croatia	1,573,302	54,603	27,293–120,114
Lithuania	1,406,261	45,298	22,644–99,644
Slovenia	949,469	29,374	14,698–64,647
Latvia	848,155	27,521	13,760–60,549
Estonia	671,861	22,031	11,013–48,459
Luxembourg	278,840	10,387	5,182–22,838
Iceland	121,760	4,575	2,290–10,077

Occupations most exposed to AI displacement by 2030 (all countries)

Occupation	Est. jobs lost (mid, all countries)
Shop Sales Assistant	3,588,821
General Office Clerk	3,160,198
Field Crop and Vegetable Growers	2,413,004
Field Crop and Vegetable Grower	1,777,163
Shopkeeper	1,729,740
General Office Clerks	1,494,106
Managing Directors and Chief Executives	1,125,777
Shopkeepers	1,075,965
Mixed Crop and Animal Producer	991,518
Mixed Crop Grower	987,201
Data Entry Clerk	954,003
Car, Taxi and Van Driver	824,249
Motorcycle Driver	737,952
Door-to-door Salesperson	597,114
Crop Farm Labourer	570,497
Livestock and Dairy Producer	563,983
Shop Sales Assistants	535,775
Data Entry Clerks	489,480
Secondary Education Teacher	477,865
Cook	463,377

How this compares to published estimates

These are not directly comparable — each measures something different. We anchor our mid scenario to the same order of magnitude, and keep the full method open.

Source	Headline	What it measures
Goldman Sachs (2023)	~300M jobs exposed; ~2.5% of US at risk of elimination	Task exposure / degradation, not net loss
WEF Future of Jobs 2025	92M displaced, 170M created (net +78M) by 2030	Employer-surveyed gross displacement & creation
McKinsey	60–70% of activities automatable	Work-activity automation potential under adoption scenarios
ILO WP140 (2023/25)	4 exposure gradients; no job-loss number	Occupational GenAI exposure (our input)
This site (mid)	71,122,534 (3.0%) across 46 countries	Exposure × automation share × realisation rate

Method in one line

jobs lost = workforce × GenAI exposure(aioe_score) × automation share × realisation rate. Realisation rates (low/mid/high) reflect how much of the theoretically-automatable exposure actually converts to job loss by 2030 — adoption is partial. Full detail on the methodology page. Download every number in the dataset (CSV).

Not a prediction. A scenario range for orientation; real outcomes depend on adoption, policy, and job creation that this figure does not net out.