What Jobs Will AI Replace? 427 Occupations

Career X-Ray · ILO data

How exposed each occupation is to generative AI, as the ILO scored it

Mean score: the average of an occupation’s task scores, from 0 to 1. The ILO counts 112 of 427 occupations as exposed, in four gradients; Gradient 4 is the highest. Occupation names are ISCO-08 titles, which keep the ILO’s international (British) spelling. Select a column head to sort.

Go straight to your job

Exposed (112)Minimal or none (315)

The 112 occupations the ILO counts as exposed, highest gradient first

Occupation (ISCO-08)	Category	Mean score
Data entry clerks 4132	Gradient 4	0.70
Typists and word processing operators 4131	Gradient 4	0.65
Accounting and bookkeeping clerks 4311	Gradient 4	0.64
Statistical, finance and insurance clerks 4312	Gradient 4	0.64
Clerical support workers not elsewhere classified 4419	Gradient 4	0.63
Securities and finance dealers and brokers 3311	Gradient 4	0.63
Financial analysts 2413	Gradient 4	0.62
Contact centre salespersons 5244	Gradient 4	0.61
Payroll clerks 4313	Gradient 4	0.61
Credit and loans officers 3312	Gradient 4	0.60
General office clerks 4110	Gradient 4	0.60
Personnel clerks 4416	Gradient 4	0.60
Web and multimedia developers 2513	Gradient 4	0.60
Translators, interpreters and other linguists 2643	Gradient 3	0.59
Bank tellers and related clerks 4211	Gradient 3	0.58
Contact centre information clerks 4222	Gradient 3	0.58
Scribes and related workers 4414	Gradient 3	0.58
Secretaries (general) 4120	Gradient 3	0.58
Applications programmers 2514	Gradient 3	0.57
Clearing and forwarding agents 3331	Gradient 3	0.57
Database designers and administrators 2521	Gradient 3	0.57
Enquiry clerks 4225	Gradient 3	0.57
Financial and investment advisers 2412	Gradient 3	0.57
Receptionists (general) 4226	Gradient 3	0.57
Statistical, mathematical and related associate professionals 3314	Gradient 3	0.57
Systems administrators 2522	Gradient 3	0.57
Mathematicians, actuaries and statisticians 2120	Gradient 3	0.56
Travel consultants and clerks 4221	Gradient 3	0.56
Advertising and marketing professionals 2431	Gradient 3	0.55
Authors and related writers 2641	Gradient 3	0.55
Economists 2631	Gradient 3	0.55
Legal secretaries 3342	Gradient 3	0.55
Librarians and related information professionals 2622	Gradient 3	0.55
Software and applications developers and analysts not elsewhere classified 2519	Gradient 3	0.55
Survey and market research interviewers 4227	Gradient 3	0.55
Administrative and executive secretaries 3343	Gradient 3	0.54
Journalists 2642	Gradient 3	0.54
Meteorologists 2112	Gradient 3	0.54
Telephone switchboard operators 4223	Gradient 3	0.54
Insurance representatives 3321	Gradient 3	0.53
Medical secretaries 3344	Gradient 3	0.53
Software developers 2512	Gradient 3	0.53
Web technicians 3514	Gradient 3	0.53
Computer network professionals 2523	Gradient 3	0.52
Medical records and health information technicians 3252	Gradient 3	0.52
Accountants 2411	Gradient 3	0.51
Coding, proof-reading and related clerks 4413	Gradient 3	0.51
Hotel receptionists 4224	Gradient 3	0.51
Information and communications technology sales professionals 2434	Gradient 3	0.51
Technical and medical sales professionals (excluding ICT) 2433	Gradient 3	0.50
Transport clerks 4323	Gradient 3	0.50
Accounting associate professionals 3313	Gradient 2	0.49
Commercial sales representatives 3322	Gradient 2	0.49
Database and network professionals not elsewhere classified 2529	Gradient 2	0.49
Government tax and excise officials 3352	Gradient 2	0.49
Graphic and multimedia designers 2166	Gradient 2	0.49
Systems analysts 2511	Gradient 2	0.49
Conference and event planners 3332	Gradient 2	0.48
Pawnbrokers and money-lenders 4213	Gradient 2	0.48
Sociologists, anthropologists and related professionals 2632	Gradient 2	0.48
Telecommunications engineers 2153	Gradient 2	0.48
Announcers on radio, television and other media 2656	Gradient 2	0.47
Archivists and curators 2621	Gradient 2	0.47
Financial and insurance services branch managers 1346	Gradient 2	0.47
Information and communications technology user support technicians 3512	Gradient 2	0.47
Information technology trainers 2356	Gradient 2	0.47
Philosophers, historians and political scientists 2633	Gradient 2	0.47
Door to door salespersons 5243	Gradient 2	0.46
Library clerks 4411	Gradient 2	0.46
Management and organization analysts 2421	Gradient 2	0.46
Bookmakers, croupiers and related gaming workers 4212	Gradient 2	0.45
Government social benefits officials 3353	Gradient 2	0.45
Personnel and careers professionals 2423	Gradient 2	0.45
Valuers and loss assessors 3315	Gradient 2	0.45
Business services agents not elsewhere classified 3339	Gradient 2	0.44
Cartographers and surveyors 2165	Gradient 2	0.44
Information and communications technology service managers 1330	Gradient 2	0.44
Production clerks 4322	Gradient 2	0.44
Retail and wholesale trade managers 1420	Gradient 2	0.44
Trade brokers 3324	Gradient 2	0.44
Computer network and systems technicians 3513	Gradient 2	0.43
Debt-collectors and related workers 4214	Gradient 2	0.43
Government licensing officials 3354	Gradient 2	0.43
Information and communications technology operations technicians 3511	Gradient 2	0.43
Office supervisors 3341	Gradient 2	0.43
Public relations professionals 2432	Gradient 2	0.43
Shop keepers 5221	Gradient 2	0.43
Business services and administration managers not elsewhere classified 1219	Gradient 2	0.42
Client information workers not elsewhere classified 4229	Gradient 2	0.42
Electronics engineers 2152	Gradient 2	0.42
Sales demonstrators 5242	Gradient 2	0.42
Mail carriers and sorting clerks 4412	Gradient 2	0.41
Sales and marketing managers 1221	Gradient 2	0.41
Training and staff development professionals 2424	Gradient 2	0.41
Biologists, botanists, zoologists and related professionals 2131	Gradient 2	0.40
Cashiers and ticket clerks 5230	Gradient 1	0.39
Education methods specialists 2351	Gradient 1	0.39
Legal and related associate professionals 3411	Gradient 1	0.39
Photographers 3431	Gradient 1	0.39
Psychologists 2634	Gradient 1	0.39
Life science technicians (excluding medical) 3141	Gradient 1	0.38
Meter readers and vending-machine collectors 9623	Gradient 1	0.38
Physicists and astronomers 2111	Gradient 1	0.38
Pre-press technicians 7321	Gradient 1	0.38
Shop sales assistants 5223	Gradient 1	0.38
Filing and copying clerks 4415	Gradient 1	0.37
Gallery, museum and library technicians 3433	Gradient 1	0.37
Hotel managers 1411	Gradient 1	0.37
Messengers, package deliverers and luggage porters 9621	Gradient 1	0.37
Medical assistants 3256	Gradient 1	0.35
Stall and market salespersons 5211	Gradient 1	0.35
Car, taxi and van drivers 8322	Gradient 1	0.28

The table is wider than the screen: swipe sideways.

Show all 112 occupations

Source: Gmyrek et al. (2025), ILO Working Paper 140, Table A1, licensed CC BY 4.0; adapted by Wild Garden (sortable, ISCO-08 titles from the data annex, two decimals, the two views). The ILO’s adaptation notice is under Sources.

Look up how exposed your occupation’s tasks are

Job title or ISCO-08 code

Example until you pick yours

Software developers (ISCO-08 2512)

0.53 out of 1

is the average of this occupation’s task scores for exposure to generative AI, on a scale from 0 to 1 where higher means more exposed.

Tick the tasks below that are yours: their average then appears here, beside the occupation’s score.

ILO category: Gradient 3, the second highest of the four gradients the ILO counts as exposed (in its words: significant exposure, high task variability).

The 8 tasks scored for this occupation, highest first. Tick the ones that are yours.

Developing and directing software testing and validation procedures0.66

Researching, analysing and evaluating requirements for software applications and operating systems0.61

Assessing, developing, upgrading and documenting maintenance procedures for operating systems, communications environments and applications software0.53

Directing software programming and development of documentation0.52

Researching, designing, and developing computer software systems0.50

Modifying existing software to correct errors, to adapt it to new hardware or to upgrade interfaces and improve performance0.50

Consulting with engineering staff to evaluate interfaces between hardware and software0.45

Consulting with customers concerning maintenance of software system0.45

What now? Pick your own occupation or US job title in the search, then tick the tasks above that are yours.

How the scores were made: AI models (GPT-4o and Gemini 1.5 Flash, averaged) predicted each task score after the ILO calibrated them on a worker survey and expert reviews; experts set 46 of the 3,265 scores. The scores are a snapshot from the start of 2025.

Spread of the task scores (standard deviation): 0.07. Mean, spread and category as printed in ILO Working Paper 140 (May 2025), Table A1; task scores from its data annex, rounded to two decimals. The average of your ticked tasks is ours, from the same task scores.

A potential, not a forecast about your job. The data describe tasks that technology could have done when the data were collected, not what your employer will do.

Your choice stays in your browser. We see none of it.

Remind me in 90 days

A calendar file with an event in 90 days, to look at your own tasks again. It is in your calendar only once you open the file.

Card to share

An image of the score above, made on your device.

Dieses Tool auf Deutsch: Welche Jobs werden durch KI ersetzt?

What the research shows: tasks, not whole jobs

Generative AI rarely takes over a whole job. It changes parts of many. In 2025 the International Labour Organization (ILO) scored the tasks of 427 occupations for exposure to generative AI and counts 112 of them as exposed, in four gradients. Clerical work leads: of the 13 occupations in the highest gradient, “the majority represent clerical roles, including data entry clerks, typists and word processing operators, accounting and bookkeeping clerks, and general office clerks.”

Across the world’s workforce, the ILO finds that “one in four jobs globally (24%) has some degree of exposure, and one in three jobs in high-income countries.” 3.3 percent of global employment is in the highest category. Since the ILO’s first index in 2023, some strongly digitized professional and technical jobs have become more exposed, while some clerical task scores went down: “While tasks such as taking meeting notes or scheduling appointments can significantly benefit from GenAI support, our 2023 scores for such tasks, which reached 0.9 in some cases, reflected an overly optimistic assessment of full automation potential.”

A high score does not mean a job disappears overnight. The ILO: “Notwithstanding these findings, we stress that such exposure does not imply the immediate automation of an entire occupation, but rather the potential for a large share of its current tasks to be performed using this technology.” That is not an all-clear either. The same paper warns: “However, this does not mean that demand for these occupations will remain stable, as the efficiencies gained from introducing technology may mean that fewer workers are needed.” And: “If efforts are not made to help such jobs evolve with GenAI and integrate new tasks, even partial automation of existing tasks could lead to a decline in overall job demand in occupations found in higher exposure gradients.”

Will AI take my job?

No index can tell you for certain. What the research measures is tasks: how much of each one generative AI could do. The ILO’s answer, in the abstract of its paper: “As most occupations consist of tasks that require human input, transformation of jobs is the most likely impact of GenAI.” As the ILO warns above, a job that is transformed can still need fewer workers, so a low score is no all-clear and a high one no verdict. The useful question is which of your own tasks are exposed: look up your occupation in the tool above, tick the tasks that are yours and compare their average with the occupation’s score.

How to read your result

The big number is your occupation’s mean score: the average of the scores of its tasks, on a scale from 0 to 1. The category under it comes from that mean and from how much the task scores differ, the standard deviation; the table below gives the ILO’s rules. Under it you see every task the ILO scored for your occupation, highest first, so you can tell which parts of the work carry the exposure.

The occupations are those of the International Standard Classification of Occupations (ISCO-08). They are broad: your own job may hold tasks that are not on the list, and spend more or less time on the ones that are. The score describes the occupation, not you.

How the scores were made

The index comes from the ILO and NASK-PIB, Poland’s National Research Institute. The paper explains: “the Research department of the ILO has partnered with NASK, the National Research Institute of the Ministry of Digital Affairs in Poland, to develop a more precise assessment of potential effects of GenAI on employment across countries.” The team first asked people about their own work: “In addition, using a representative sample from the 29,753 tasks in the Polish occupational classification system and a survey of 1,640 people in Poland, we captured the opinions of each 1-digit ISCO-08 group on the potential for task automation in the job categories in which they are currently employed (52,558 data points).” “We then compared this input with an expert survey and conducted subsequent rounds of revisions among a smaller international expert group for the tasks with the highest degree of disagreement.”

The scores you see were not rated by the survey’s respondents. The ILO used the adjusted survey scores to build an AI assistant: “Once the prompt adjustments deliver satisfactory performance, we use the same system to predict automation scores for 3,265 tasks in ISCO-08 (as in Gmyrek et al., 2023).” Its models were OpenAI’s GPT-4o and Google’s Gemini 1.5 Flash. In the published data, each task score is the mean of the two models’ predictions, except 46 of the 3,265 that experts set by reconciliation. The paper also describes a manual review of 364 tasks whose scores had moved most since 2023, 64 of them settled in a group discussion; how those relate to the 46 reconciled scores of the data annex, it does not say. The ILO dates the whole result to the start of 2025.

To sort occupations, the ILO uses “occupational mean and standard deviation (SD) of task-level scores pertaining to a given 4-digit occupation.”

Exposure categories: the rules of Table 5 (μ mean, σ standard deviation of the task scores)

The table is wider than the screen: swipe sideways.

The categories are the ILO’s. For 4 occupations the printed mean and standard deviation add up to exactly 0.40 although the category is minimal exposure, which needs more than 0.4: the ILO set the category before rounding. Recomputed from the data annex’s task scores alone, 12 occupations would fall in another category. We show the categories as the ILO prints them and say so at each of these results. Source: ILO Working Paper 140, Table 5 (rules, quoted as printed) and Table A1 (our tally of its rows per category); task scores from the paper’s data annex, rounded to two decimals.

Three examples

Data entry clerks (ISCO-08 4132). Mean 0.70, standard deviation 0.03. The mean is at least 0.6, and 0.70 minus 0.03 is 0.67, at least 0.5: Gradient 4. The ILO’s reading of that category: “High and consistent GenAI exposure across tasks within the occupation. Most current tasks in these jobs have a high potential of automation, with little variability in task-level exposure.”

Software developers (ISCO-08 2512). Mean 0.53, standard deviation 0.07. The mean is at least 0.5 and below 0.6, and 0.53 plus 0.07 is 0.60, at least 0.5: Gradient 3. “Above-moderate occupational exposure: even though some tasks remain less exposed, the overall potential of automation of the current tasks with GenAI is growing in these occupations.”

Nursing professionals (ISCO-08 2221). Mean 0.25, standard deviation 0.13. Mean plus standard deviation is 0.38, below 0.5 and not above 0.4, so none of the exposed gradients and not minimal exposure: not exposed. “Occupations where most tasks remain relatively unaffected by GenAI, with low task variability and a stable low mean exposure score.”

Another measure: the US AI exposure categories

In August 2026 the US Bureau of Labor Statistics added AI exposure categories to its projections. It combined five outside measures and sorted 831 US occupations into four groups of relative exposure: low, moderate, high and very high. Relative is the key word: “The exposure category is relative to other occupations. It does not imply an absolute ‘high’ or ‘low’ level of exposure.” The BLS ranks occupations against each other; the ILO scores each task on its own scale from 0 to 1. So the two can disagree. The BLS puts registered nurses, secondary school teachers and paralegals in its group “High.” In the ILO’s index, nursing professionals average 0.25 and secondary education teachers 0.30, both not exposed, and legal and related associate professionals, where the crosswalk files paralegals, 0.39, Gradient 1. Both measure exposure, not a forecast of jobs gained or lost; the BLS says so of its categories too.

How the search matches US job titles

The ILO scores ISCO-08 occupations, which few people in the US use. So the search also knows 16,699 US job titles from the O*NET 31.0 database, in 840 US occupations, and maps each to ISCO-08 in three steps, each published by a US agency or, for O*NET, by the O*NET Center that the US Department of Labor sponsors: the O*NET occupation to the 2018 Standard Occupational Classification (SOC), the 2018 SOC to the 2010 SOC, and the 2010 SOC to ISCO-08 with the crosswalk of the Bureau of Labor Statistics (BLS). Every US-title result says how it got there: O*NET lists “Registered Nurse” under the US occupation Registered Nurses (SOC 29-1141), which the BLS crosswalk links to ISCO-08 2221, nursing professionals.

Some US occupations link to several ISCO-08 occupations. The BLS explains its marking: “Asterisks (*) in the column ‘part’ indicate that the ISCO code matched to more than one SOC code.” Which of them most US workers in the occupation hold is our judgment: we reviewed each such occupation by hand, show that one first and list the others to choose from; where no one of them clearly covers most workers, the result asks you to pick. When the same US title stands under several US occupations, the one whose own name shares your words comes first, then the larger occupation by US employment (BLS, 2025). Words a job ad adds, such as Senior, Lead, full-time, a shift or the pay, are left out, as is what follows a dash when no title carries it, and a four-digit number is read as an ISCO-08 code. The BLS kept its crosswalk lean on purpose: “The principle of parsimony was used to make the crosswalk more straightforward and therefore more useful.” Armed forces occupations are not in the ILO table: “Armed Forces Occupations have been excluded due to the lack of specific tasks.”

Working in Germany?

For Germany there is a separate measure: the IAB’s share of an occupation’s core tasks that computers could fully automate, with the technology of 2022, next to the median pay of the occupation group from Germany’s Federal Employment Agency. Our German page shows it for the occupation group (Berufsgattung) of a German job title, where the German tables have a value: Welche Jobs werden durch KI ersetzt?. The two measures are not on one scale.

What the scores do not tell you

First: these are estimates. AI models predicted the task scores from a base of survey answers and expert rounds, and the ILO has already revised scores it found too high. Second: an occupation is not your job. ISCO-08 groups many jobs under one title, and a US title reaches it through a crosswalk built for statistics, not for one person. Third: exposure is not a decision. Whether tasks get automated depends on cost, quality, law and the people who do the work, and none of that is in a score.

Frequently asked questions

What jobs will AI replace?

Very few jobs as a whole; many jobs in part. The ILO scored the tasks of 427 occupations for exposure to generative AI and counts 112 as exposed. The highest mean scores belong to clerical work: data entry clerks (0.70), typists and word processing operators (0.65), accounting and bookkeeping clerks (0.64), statistical, finance and insurance clerks (0.64). Even there the ILO speaks of tasks, not of whole occupations vanishing at once, and it warns that fewer workers may be needed (see below).

Which jobs will AI replace first, and by 2030?

No source here gives a list of jobs with dates for AI alone. The nearest is the US Bureau of Labor Statistics. Its newest projections run from 2025 to 2035 and expect total US employment to grow 3.5 percent. Where it expects AI to reduce the need for workers, it says so; for claims adjusters it writes: “Productivity change - share decreases as computer software and artificial intelligence (AI) will make claims adjustment processes more efficient, allowing adjusters to focus on more complex claims.” It projects employment of claims adjusters, examiners and investigators to fall 5.5 percent, of auto damage insurance appraisers 8.7 percent and of data entry keyers 25.5 percent, while software developers grow 10.2 percent. Its projections before, for 2023 to 2033, already did this: “The 2023–33 BLS employment projections incorporate AI-related impacts for several occupations for which high exposure to automation is deemed likely.” And the BLS is plain about the limits: “The timing and scale of many potential impacts of GenAI are too uncertain to be reflected in BLS projections.” The ILO adds where exposure is highest today: “Clerical occupations continue to have the highest exposure levels.”

Is my job safe from AI?

No index can promise that, for the next five years or any other span, and this one does not try. It shows how exposed the tasks of your occupation are, as the ILO scored them, and the ILO is clear about the limits: “In that sense, our estimates present a static snapshot of the exposure of existing occupations to GenAI at the start of 2025.” And: “Finally, the estimates do not account for new jobs that may be created, nor the technological advances that could potentially automate more tasks in the future.” Whether your employer automates your tasks is a different question.

Which jobs are least exposed to AI?

The table above has a view for them, “Minimal or none”: the 315 occupations below the ILO’s four gradients, 231 not exposed and 84 with minimal exposure, lowest mean first. 12 occupations share the lowest mean score, 0.09, among them bricklayers and related workers, building construction labourers, building frame and related trades workers not elsewhere classified. Least exposed describes today’s tasks as the ILO scored them for generative AI; it is not a promise.

What does the exposure score mean?

It is the average of the scores of an occupation’s tasks, from 0 for a task current generative AI cannot perform to 1 for one it could perform entirely. AI models predicted each task’s score after the ILO calibrated them on a worker survey and expert reviews. The ILO then sorts occupations by their mean and the spread of their task scores into exposure categories.

Does a high score mean I will lose my job?

Not by itself. The ILO: “Notwithstanding these findings, we stress that such exposure does not imply the immediate automation of an entire occupation, but rather the potential for a large share of its current tasks to be performed using this technology.” That is not an all-clear either: “However, this does not mean that demand for these occupations will remain stable, as the efficiencies gained from introducing technology may mean that fewer workers are needed.”

How were the scores made, and how accurate are they?

They are estimates. The ILO surveyed 1,640 workers in Poland and ran expert reviews, then built an AI assistant that predicts a score for each task. In the published data, each score is the mean of two models, GPT-4o and Gemini 1.5 Flash, except 46 that experts set by reconciliation. The paper tests the assistant on survey tasks it had not seen: “These tasks were split into an 80-20 proportion, with 80% used for training and 20% reserved for validation.” The result: “Performance was measured using Root Mean Square Error (RMSE) and Mean Absolute Error (MAE), with the adjusted scores showing better agreement (MAE = 0.0494, RMSE = 0.0724) compared to the predicted scores (MAE = 0.0684, RMSE = 0.0963).” An average error of 0.0494 is about five hundredths on the 0-to-1 scale. The two models agree closely, at a correlation of 0.96, and predictions from ISCO-08 tasks and from Polish tasks correlate at 0.92. The ILO has also revised scores it found too high in its first index.

What can I do with my result?

Use it to look at your own work, not as a verdict on you. Which of your tasks look like the high-scoring ones, and which need your judgment, your contact with people, your experience? The second group is where you can build. That takes time and practice, and it is the part you control.

Privacy

The search and your result run in your browser. The ILO data and the US job titles are part of this page; what you type or choose does not leave your device, and the tool writes nothing into the link. The share image and the calendar file are made by your browser; whether and with whom you share them is up to you. Your choice stays in your browser. We see none of it. More in our privacy policy.

Sources and licenses

All sources retrieved from October 21 to 25, 2026 and archived with the project.

• Gmyrek, P., Berg, J., Kamiński, K., Konopczyński, F., Ładna, A., Nafradi, B., Rosłaniec, K., Troszyński, M. 2025. Generative AI and Jobs: A Refined Global Index of Occupational Exposure. ILO Working Paper 140. Geneva: International Labour Office. © ILO. doi.org/10.54394/HETP0387, PDF. Licensed under CC BY 4.0. The paper is co-authored by researchers of the ILO and of NASK-PIB.
• Task-level scores: Gmyrek et al. (2025), the paper’s data annex, github.com/pgmyrek/2025_GenAI_scores_ISCO08, licensed under CC BY 4.0 (LICENSE).
• Changes we made to the ILO material: the table of exposed occupations and every result are adapted from Table A1 and the annex (sortable, ISCO-08 titles from the annex, two decimals); task scores rounded to two decimals; in 63 places we repaired characters the annex stores in a broken encoding (for example “students’ performance”). As the ILO license requires: “This is an adaptation of a copyrighted work of the International Labour Organization (ILO). This adaptation has not been prepared, reviewed or endorsed by the ILO and should not be considered an official ILO adaptation. The ILO disclaims all responsibility for its content and accuracy. Responsibility rests solely with the author(s) of the adaptation.”
• US job titles: This page includes information from the O*NET 31.0 Database by the U.S. Department of Labor, Employment and Training Administration (USDOL/ETA). Used under the CC BY 4.0 license. O*NET® is a trademark of USDOL/ETA. Wild Garden has modified all or some of this information. USDOL/ETA has not approved, endorsed, or tested these modifications. License: CC BY 4.0; files: O*NET 31.0 Database (Occupation Data, Sample of Reported Titles, Job Titles) and the O*NET-SOC 2019 to 2018 SOC crosswalk. Changes: titles grouped by occupation, duplicates removed, mapped to ISCO-08.
• U.S. Bureau of Labor Statistics: ISCO-08 to 2010 SOC crosswalk (August 2012, updated June 2015) and 2010 to 2018 SOC crosswalk; public domain, cited as the BLS asks.
• U.S. Bureau of Labor Statistics, Employment Projections 2025–35: Table 1.2, occupational projections and worker characteristics, and Table 1.12, factors affecting occupational utilization (employment in 2025, used to order equally good search options; the projected changes and one factor quoted in the FAQ); public domain.
• U.S. Bureau of Labor Statistics: Employment Projections: 2025-2035 Summary, news release USDL-26-1422, August 27, 2026, and AI exposure categories with its data file; public domain (quoted).
• Machovec, Christine; Rieley, Michael J.; Rolen, Emily (2025): Incorporating AI impacts in BLS employment projections: occupational case studies, Monthly Labor Review, February 2025, U.S. Bureau of Labor Statistics; public domain (quoted).
• Every BLS file above was retrieved as archived by the Internet Archive, because bls.gov refuses automated requests.

Published October 25, 2026

The ILO index scores tasks for generative AI on a 0 to 1 scale; exposure is not a forecast of job losses. Scores describe tasks of an occupation as the ILO defines it (ISCO-08), not your own job; they are a snapshot from the start of 2025, predicted by AI models the ILO calibrated on a worker survey and expert reviews. The content is general information and does not replace individual medical, legal, tax or financial advice.

More from the garden

All calculators and tools are under Tools. To build a new skill into your week, the habit formation calculator turns an intention into a routine. If you are weighing a change, a buffer gives you room: the emergency fund calculator shows how large yours should be. And the wheel of life puts work next to the other areas of your life.

Content and tools on Wild Garden are published by the Wild Garden Editorial Team. How we research, check and correct errors is set out in our editorial standards.