Topic · AI and jobs

How AI is changing the job market: what the major reports say

By · Updated

In short

The major reports agree on the direction but not on doom. The World Economic Forum expects 170 million jobs to be created and 92 million displaced by 2030, a net gain of 78 million, with 39% of key skills changing. The IMF estimates that almost 40% of jobs worldwide, and about 60% in advanced economies, are exposed to AI; the ILO finds one in four workers in an occupation with some exposure to generative AI and expects most of those jobs to be transformed rather than eliminated. The first measurable losses are among young people: in US payroll data, employment of 22- to 25-year-olds in the most exposed occupations fell 13% relative to others. Pay is moving the other way for people who use AI well: PwC finds a 62% wage premium for jobs that ask for AI skills.

Key facts

Jobs to 2030 (WEF)
170 million created, 92 million displaced: net +78 million
Jobs exposed to AI (IMF)
Almost 40% worldwide, about 60% in advanced economies
Exposed to generative AI (ILO)
1 in 4 workers; 3.3% in the highest exposure group
Young workers (Stanford)
−13% relative employment for ages 22 to 25 in the most exposed jobs
AI skills pay (PwC)
62% wage premium in job ads that ask for AI skills
Skills changing by 2030 (WEF)
39% of the key skills employers need

The reports at a glance

What six major studies found about AI and work
StudyMain findingBasis
World Economic Forum, Future of Jobs 2025170 million jobs created and 92 million displaced by 2030; 39% of key skills changeSurvey of more than 1,000 large employers covering over 14 million workers, January 2025
IMF, Gen-AI and the Future of WorkAlmost 40% of global employment exposed to AI; about 60% in advanced economies, 40% in emerging markets, 26% in low-income countriesStaff Discussion Note, January 2024
ILO, Generative AI and jobs1 in 4 workers in an occupation with some generative AI exposure; 34% of employment in high-income countries against 11% in low-income onesRefined global index and 2025 update, May 2025
Stanford, Canaries in the Coal Mine?13% relative fall in employment of 22- to 25-year-olds in the most AI-exposed occupationsUS payroll records, August 2025
PwC, Global AI Jobs Barometer 202662% wage premium for AI skills; headcount up 52% at the most AI-exposed companies against 36% at the least, 2018 to 2025More than one billion job ads in 27 countries, June 2026
Anthropic Economic IndexComputer and mathematical tasks, mostly coding, are 35% of Claude.ai conversationsSample of February 2026, report of March 2026

Exposure means AI can perform or assist with a large share of an occupation’s tasks. It is not a forecast of job losses.

Pay is where these shifts end up. See what any salary leaves after tax in nine countries with the free calculator. Open the take-home pay calculator

Watch: AI and jobs in one story

Narrated animation with captions, 1 minute 25 seconds. Watch on YouTube.

AI and jobs

Exposure is not the same as job loss

The headline numbers measure exposure: how many of a job’s tasks AI could do or help with. The IMF puts almost 40% of global employment in that group, rising to about 60% in advanced economies because they have more office and professional work. It then splits the exposed jobs in two. In advanced economies about 27% of all employment is exposed but likely to be complemented by AI, which can raise productivity and pay, and about 33% is exposed with less complementarity, where AI is more likely to replace tasks.

The ILO’s index, refined in 2025, is narrower and finds one in four workers in an occupation with some exposure to generative AI, but only 3.3% in the highest exposure group. Clerical jobs are the most exposed, and women are more exposed than men: 4.7% of female employment is in the top group against 2.4% of male employment. The ILO’s conclusion is that because people are still needed to check and direct the work, most jobs will be transformed rather than made redundant.

Where the first effects show: young workers

The clearest evidence of job losses so far is at the start of careers. Using payroll records from the largest US payroll software provider, Stanford economists Erik Brynjolfsson, Bharat Chandar and Ruyu Chen found that employment of 22- to 25-year-olds in the most AI-exposed occupations fell 13% relative to others after generative AI spread. Employment of older workers in the same occupations held steady or grew.

The losses were concentrated where AI automates work rather than assists with it, and they came through fewer hires rather than lower pay. PwC’s 2026 barometer adds a nuance from US job ads: entry-level roles exposed to AI are seven times more likely than other entry-level roles to ask for senior-level skills such as leadership or creativity, and their openings have risen 35% since 2019 while other entry-level openings fell 10%. The entry rung is not disappearing; it is asking for more.

Which jobs grow

The World Economic Forum’s survey of more than 1,000 large employers expects the fastest growth, in percentage terms, for big data specialists, fintech engineers and AI and machine learning specialists. In absolute numbers the largest gains are elsewhere: farmworkers, with 34 million more jobs by 2030, then delivery drivers, software developers, building construction workers and shop salespeople, followed by care work such as nursing and social work.

The same survey expects 39% of the key skills employers need to change by 2030, down from 44% in its 2023 edition. AI and big data, networks and cybersecurity and technological literacy grow fastest, but so do creative thinking, resilience and flexibility, curiosity and lifelong learning, and leadership.

What it means for pay

PwC’s analysis of more than a billion job ads in 27 countries finds that jobs asking for AI skills pay 62% more on average than similar jobs that do not, up from 57% a year earlier, ranging from 118% in consumer markets to 16% in government. Between 2018 and 2025, the companies most exposed to AI grew their headcount by 52% and wages by 24%, against 36% and 17% at the least exposed.

PwC describes a labour market splitting into two tracks. In professionalised roles, such as radiologists and recruiters, AI takes the routine tasks and human judgement becomes more valuable; these roles grow twice as fast and their pay 42% faster. In democratised roles, such as IT service managers and medical secretaries, AI makes the work easier for non-specialists and growth is slower.

What it means in Spain and the UAE

Both are high-income economies, the group the ILO finds most exposed, with 34% of employment against 11% in low-income countries. Spain’s large services and administrative workforce sits in the categories with the most exposure, and its high youth unemployment makes the Stanford finding on entry-level hiring the one to watch. The UAE, where most of the workforce is expatriate and hired for specific roles, adjusts mainly through who gets hired and on what package.

The reports are global or American; none measures Spain or the UAE directly, so country effects are an inference, not a finding. What the data supports in both places is the same practical point: the premium goes to people who combine AI with judgement, client work and responsibility for results.

Borja’s view

Borja spent years designing pay structures as a compensation and benefits consultant, and his reading of these reports is that pay follows scarcity. When AI makes routine work cheap, the scarce things become judgement, trust and the ability to take responsibility for a result, and that is where the premium goes. The next step is machines that pay each other for work, which needs a form of money that software can hold and send without a bank, the argument at the heart of Why Bitcoin Had to Exist.

Questions

Will AI destroy more jobs than it creates?

Not according to the World Economic Forum’s employer survey, which expects 170 million jobs created and 92 million displaced by 2030, a net gain of 78 million. The mix of jobs and skills changes a lot more than the total.

How many jobs are exposed to AI?

The IMF estimates almost 40% of global employment and about 60% in advanced economies. The ILO’s stricter index puts one in four workers in an occupation with some generative AI exposure, and 3.3% in the highest exposure group.

Is AI already reducing jobs for young people?

In the United States, Stanford researchers found a 13% relative fall in employment for 22- to 25-year-olds in the most AI-exposed occupations, mainly through fewer hires. Older workers in the same jobs were not affected in the same way.

Which jobs are most exposed to AI?

Clerical and administrative jobs have the highest exposure in the ILO’s index. Media and web roles have become more exposed as AI generates images, voice and video.

Do AI skills pay more?

Yes. PwC’s 2026 barometer finds that job ads asking for AI skills offer 62% more on average than comparable ads that do not.

What skills grow most because of AI?

AI and big data, networks and cybersecurity and technological literacy, according to the World Economic Forum, together with creative thinking, resilience, curiosity and lifelong learning, and leadership.

Sources

  1. World Economic Forum, Future of Jobs Report 2025: the jobs of the future and the skills you need to get them
  2. IMF, Gen-AI: Artificial Intelligence and the Future of Work, Staff Discussion Note SDN/2024/001
  3. ILO, Generative AI and jobs: a refined global index of occupational exposure, Working Paper 140
  4. ILO, Generative AI and jobs: 2025 update, research brief
  5. Stanford SIEPR, Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence
  6. PwC, 2026 Global AI Jobs Barometer, press release (15 June 2026)
  7. Anthropic Economic Index, Learning curves (March 2026)

Updated:

Free tools

Calculators built on the official rules, with their sources. Free, no sign-up.

See all free tools →

More books

Guides

Bitcoin from zero: a free course in six short lessonsA free course in six short lessons: what Bitcoin is, transactions, wallets, safety, scams and tax in Spain.What is Bitcoin, and why does it exist?What it is, who made it, and why a money with fixed rules had to exist.What are stablecoins, and why do they matter?Digital dollars: how they hold their value, why they matter and the rules arriving.Money and pay: why money loses value and how pay is changingInflation, total rewards, pay transparency and what digital money changes for HR.Web3 and the business of sportFan tokens, crypto sponsorship and prediction markets: how money is changing sport.Can you be paid your salary in crypto or stablecoins?Can your salary be paid in crypto or stablecoins? The rules in the UAE and Spain.The best books to understand Bitcoin, for beginnersThe best books to understand Bitcoin, from first steps to the technical detail.The best books on stablecoinsThe best books and reports to understand stablecoins, from inclusion to regulation.Which banks are launching stablecoins, and what does it mean for you?Qivalis, Santander, BBVA, CaixaBank: which banks are launching stablecoins, and what it means for customers.Books on Web3 and sport: what to read firstThe few books worth reading on Web3, fan tokens and the business of sport, and what each one is for.Stablecoins in the UAE: what is allowed, who issues them and what it means for youWhat the UAE allows, who regulates it, and the first dirham stablecoins: AE Coin, DDSC and RAKBank.Open lists in Spain: how Spaniards vote today, and what would changeHow Spaniards vote for Congress and the Senate today, and what open lists would change.How does the D’Hondt method work? Spain’s seat formula, step by stepHow Spain turns votes into seats, with a worked example, and why the province matters more than the formula.How to vote from abroad in Spain’s 29 November 2026 electionHow Spaniards living abroad vote on 29 November 2026: the consulate urn, postal vote and deadlines.How many votes does a seat cost in Spain? Province by province, 29 November 2026How many voters stand behind each seat in Congress, province by province, for 29 November 2026.Postal vote for 29 November 2026: request it by 19 NovemberHow to vote by post from Spain on 29 November 2026: request by 19 November, send by 25 November.Polling station duty on 29 November 2026: pay, hours and time offCalled to sit on a polling station on 29 November 2026? Pay, hours, time off work and what to do if you cannot go.Can stablecoins make sending money home cheaper?Do stablecoins make sending money home cheaper? What the World Bank and IMF data show.How pay works in the UAE: basic salary, allowances and gratuityBasic salary, allowances, gratuity, no income tax: what a UAE pay package really means.Paying employees in stablecoins: what employers need to know in the UAE and SpainWhat employers in the UAE and Spain can and cannot do when they pay staff in stablecoins.Take-home pay index 2026: how much of your salary you keep in nine countriesHow much of a $40,000 to $250,000 salary you keep in nine countries, under 2026 rules.Spain vs Dubai salary 2026: what you keep, and the Spanish salary that matches DubaiWhat the same salary leaves in Spain and in Dubai, and the Spanish salary that matches a Dubai one.Will Bitcoin be the money of AI agents? How machines pay, and what is still missingAI agents are starting to pay for things. Cards, stablecoins and Bitcoin compared as money for machines.Why does Bitcoin go up and down so much?Fixed supply, big funds, interest rates and leverage: why Bitcoin’s price swings so much.How a Bitcoin transaction works, step by stepWhat happens between pressing send and the payment being final: the queue, the fee, the block and the confirmations.What is a Bitcoin wallet? Keys, seed phrase and custodyA wallet does not hold coins: it holds keys. What the seed phrase is, and custody by an exchange versus your own.How to keep bitcoin safe: six habits for beginnersSix habits that protect bitcoin: the seed phrase on paper, never online, small tests, updates and a plan for your family.Bitcoin scams: the most common ones and how to spot themThe most common bitcoin scams, what they have in common, and how to check a platform before you send money.How bitcoin is taxed in Spain in 2026When bitcoin is taxed in Spain, the 19% to 30% scale, crypto-to-crypto swaps, FIFO and the Modelo 721.Token compensation explained: grants, vesting, lockups and taxHow token grants, vesting, cliffs and lockups work, how they differ from shares, and when they are taxed.Jobs in stablecoins and digital money: the roles, the skills and the talentThe roles stablecoins and tokenisation create in banks, issuers and corporates, the skills they need and where the talent comes from.Paid in a dollar stablecoin: who carries the exchange-rate risk?If part of the pay is a dollar stablecoin, who carries the exchange-rate risk: the employer or the employee? Four designs, with examples.Financial literacy at work: why digital money makes it HR’s jobOnly a third of adults reach the OECD’s minimum level of financial literacy. Why digital money makes it an HR question, and what a basic programme covers.
Portrait of Borja Burguillos

Borja Burguillos

Borja writes for people who want to understand money and incentives without the jargon. Born in Madrid and based in Dubai since 2012, he has advised companies on people and pay, built tech startups, and co-chairs the People & Organisation Committee at the Middle East Stablecoin Association.

WorldatWork Total Rewards and CAIA certified, CFA Investment Foundations, executive programmes at Wharton and IE Business School.