AI Jobs Statistics 2026: Demand, Skills, Salaries & Workforce Impact
In 2025, about 2.5% of U.S. job postings listed an AI skill, according to Lightcast data published in the 2026 Stanford AI Index. That was up 55% in one year. By Indeed's wider keyword count, about 6.7% of U.S. postings mentioned AI at the end of August 2026. Both numbers are real, and they measure different things. PwC found that jobs requiring AI skills grew almost eight times faster than the overall jobs market. A Stanford-led study of payroll records found that young workers in the most exposed jobs were losing ground. The 2026 evidence does not show a boom or a bust. It shows uneven change, and the details decide what it means for you.
TL;DR
Demand is rising from a small base. In U.S. data for 2025, 2.5% (Lightcast) to 2.8% (PwC) of postings required or listed AI skills, up 55% to 66% in a year.
The pay premium is real but easy to misread. PwC reports a 62% average advertised-wage premium for AI skills across 27 countries. It is not an average AI salary, and it does not control for education, experience or location.
Skills spread beyond engineering. Python leads technical demand, agentic AI mentions rose from 0.06% to 0.23% of U.S. postings, and 51% of AI-skill postings were outside IT roles in 2024 data.
Entry-level evidence is mixed. One payroll study shows a 13% relative employment decline for ages 22 to 25 in highly exposed jobs. PwC finds AI-exposed entry-level roles growing, but demanding senior-level skills.
Forecasts need care. WEF's 170 million jobs created and 92 million displaced come from all trends combined. Its AI-specific estimate is 11 million created and 9 million displaced by 2030.
AI jobs statistics for 2026 show fast growth from a small base. About 2.5% to 2.8% of U.S. job postings required AI skills in 2025, and PwC found a 62% average advertised wage premium. Evidence on job loss is mixed: entry-level data shows pressure, but exposure to AI does not mean replacement.
Table of contents
AI Jobs Statistics 2026: Key Numbers at a Glance
The most reliable headline is this: AI skills are in more job postings than ever, pay more in advertised wages, and are changing how jobs are built, while evidence of mass job loss remains limited and contested. The table lists the main findings. Each row measures something different, so do not add or compare the rows directly.
Metric | Latest finding | Geography and period | Source |
|---|---|---|---|
Postings requiring AI skills | 2.8% of postings; more than 1.12 million postings, up 66% | U.S., 2025 data (published June 2026) | |
Postings listing an AI skill | 2.5% of postings, up 55% from 2024 | U.S., 2025 data (published April 2026) | |
Postings mentioning AI | 6.74% on Aug. 31, 2026, versus 3.44% a year earlier | U.S., Indeed postings | |
Growth in AI-specialist postings | 68.9% rise, versus 8.6% for all postings | 27 countries, 2024 to 2025 | |
AI-skill wage premium | 62% average advertised-wage premium (sector range 16% to 118%) | 27 countries, 2025 data | |
Jobs created and displaced by AI (forecast) | 11 million created, 9 million displaced | Global, 2025 to 2030 | |
Data scientist outlook (projection) | About 35% growth; 275,600 jobs in 2025 | U.S., 2025 to 2035 | |
Workers in jobs with some generative-AI exposure | About 1 in 4 workers; 3.3% in the highest tier | Global, modeled | |
Early-career employment in exposed jobs | 13% relative decline for ages 22 to 25 | U.S. payroll data (paper published August 2025) |
Three patterns stand out. First, the share of postings that involve AI is still small in the strictest measures and larger in looser ones. Second, growth rates are high mostly because the starting point is low. Third, the evidence on harm is concentrated in specific groups, such as early-career workers and clerical roles, rather than spread evenly across the economy. The sections below explain each pattern and its limits.
How to Read AI Employment Statistics Without Being Misled
AI employment statistics differ because they count different things. A posting that lists an AI skill, a job that exposes tasks to AI, and a worker who loses a job to AI are three separate facts. Mixing them up is the most common error in coverage of this topic.
Start with the vocabulary:
AI job: There is no single official definition, and we found no official worldwide count of "AI jobs." Use the term loosely at best.
Posting that requires or lists AI skills: A job ad that names an AI skill. This is a demand signal, not a headcount.
Posting that mentions AI: A looser keyword match, which includes ads where AI is only background.
AI-enabled job: A job where workers use AI tools but the title may not mention AI.
AI-exposed occupation: An occupation where AI could perform some tasks. Exposure is a model estimate, not an outcome.
Augmentation versus automation: Augmentation helps a worker do tasks. Automation does the tasks instead.
Displacement: A worker actually loses a job. Predicted displacement is a forecast, and observed displacement is rarer and harder to prove.
The table shows why sources disagree.
Data type | Example in this article | What it cannot tell you |
|---|---|---|
Job advertisements | Lightcast, PwC, Indeed | How many people hold those jobs |
Employer surveys | WEF Future of Jobs | What employers actually do |
Worker or member profiles | Jobs outside LinkedIn's membership | |
Employment records | ADP payroll study, BLS | Why employment changed |
Exposure models | ILO | Whether any job will change or vanish |
Advertised pay | PwC, Lightcast, Indeed | What workers actually earn |
Productivity data | PwC company analysis | Whether AI caused the gains |
Forecasts | WEF, BLS projections | What will happen |
Two further rules help. First, separate publication year from data year. PwC's 2026 report mostly analyzes 2025 job postings, and the Stanford AI Index 2026 reports 2025 data. Second, do not read correlation as cause. Indeed's July 2026 analysis states that its findings do not prove causation. The same period also included high interest rates, slower hiring after the pandemic boom, technology-sector cuts and remote-work shifts. Indeed reported that overall U.S. software-development postings were still about 27.5% below pre-pandemic levels, so AI is only one of several forces acting on that market.
Where we could not open a primary page during research, we say so and flag figures taken from published summaries.
How Strong Is AI Job Demand in 2026?
AI skill demand is growing quickly in 2026, but it is still a minority of all postings. In U.S. data for 2025, Lightcast found 2.5% of postings listed an AI skill. That was up 55% from 2024, and nearly 300% higher than a decade earlier, according to the Stanford AI Index summary. PwC's separate U.S. count found 2.8% of postings required AI skills, which was more than 1.12 million postings, a record.
Share and volume tell different stories. A rising share means AI skills are taking a bigger slice of hiring. A rising count could simply reflect a bigger market. In Canada, for example, PwC found about 69,000 more postings requiring AI skills in 2025, and the share rose from 1.6% to 3.5%. Both measures went up, which is stronger evidence than either alone.
The newest data suggests acceleration through 2026. Indeed's tracker shows the U.S. share of postings that mention AI roughly doubled in a year, from 3.44% to 6.74% by August 31, 2026. That series counts keyword mentions, so part of the rise may reflect employers adding AI language to ads for ordinary jobs. It confirms direction, not the number of AI roles.
The mix of demand is shifting too:
Generative and agentic AI: Mentions of agentic AI skills rose from 0.06% of U.S. postings in 2024 to 0.23% in 2025, about 90,000 postings, per Lightcast. That is a 280% jump, but still a very small share.
AI literacy: LinkedIn reported that U.S. jobs requiring AI literacy grew 70% year over year in its January 2026 report.
Beyond engineering: Lightcast's 2024 data, published July 2025, found 51% of postings requiring AI skills were outside IT and computer science occupations.
"Booming" is therefore only fair in a narrow sense. Demand for AI skills is climbing fast. Demand for AI specialists as a share of all work remains small. For wider context, see our guides on AI adoption statistics and artificial intelligence statistics.
AI Job Demand by Country and Region
Singapore leads the Lightcast comparison for AI-skill posting share, and the United States sits well below several smaller economies. No single dataset covers every country the same way, so the table groups figures by what they measure.
Market | Latest figure | Measure, period and source |
|---|---|---|
Singapore | 4.77% | Postings listing AI skills, 2025 (Lightcast / Stanford) |
Hong Kong | 3.5% | Same measure and source |
Luxembourg | 3.4% | Same measure and source |
Spain | 3.3% | Same measure and source |
United States | 2.6% | Same measure and source (the same report cites 2.5% elsewhere) |
Canada | 3.5% | Postings requiring AI skills, 2025, up from 1.6% (PwC Canada) |
Australia | 41,000 postings | Postings seeking AI skills in 2025, up from 20,000 (PwC Australia) |
Ireland | 14.89% | Postings mentioning AI, Aug. 31, 2026; 9.29% a year earlier (Indeed AI Tracker) |
United Kingdom | 9.66% | Same measure; 4.90% a year earlier |
Australia | 7.35% | Same measure; 4.85% a year earlier |
United States | 6.74% | Same measure; 3.44% a year earlier |
Germany | 5.63% | Same measure; 3.60% a year earlier |
Italy | 5.02% | Same measure; 3.64% a year earlier |
France | 4.05% | Same measure; 2.55% a year earlier |
Netherlands | 3.56% | Same measure; 2.29% a year earlier |
Lightcast and PwC count postings that list or require AI skills, so their numbers are lower. Indeed counts postings that mention AI, so its numbers are higher. Never compare a Lightcast figure for one country with an Indeed figure for another. Within the Indeed series, every country shown rose over the year, which supports a broad upward trend.
We left Canada out of the Indeed rows. The tracker file shows a much larger jump for Canada than for any other country, and we could not tie it to a published explanation. Treat it as unverified.
PwC's separate Singapore report analyzed about 1.6 million postings from 2025. It found that 82% of AI-related postings were for AI user roles and 18% for AI developer roles (PwC Singapore, June 2026). Within the U.S., Lightcast reported the highest AI share in Washington, D.C., while California leads in total volume.
India is among PwC's 27 countries. We did not find India-specific figures from an official or primary source that we could verify, so we do not report one. The same applies to the broader European Union, where comparable official AI-posting statistics were thin. If your decision depends on a market not shown, check that country's statistics agency and ask what the measure counts.
Fastest-Growing AI and AI-Enabled Jobs in 2026
The fastest-growing AI roles are growing at high percentages, but most start from small numbers. Official statistics cover broader occupations such as data scientist, not narrow titles such as "AI engineer," so the table mixes two kinds of evidence.
Role | Latest signal | Geography, period and source |
|---|---|---|
AI engineer | About 13 times more roles since 2023; roughly 298,000 new roles | Global, 2023 to 2025, LinkedIn member data (LinkedIn report) |
Forward-deployed engineer | About 42 times growth; roughly 49,000 new roles | Same source and period |
Data annotator | About 774,000 new roles | Same source and period |
AI/machine-learning specialist | Among the fastest-growing jobs by percentage | Global, 2025 to 2030 employer survey (WEF) |
Data scientists | About 35% growth; 275,600 jobs in 2025 | U.S. projection, 2025 to 2035 (BLS) |
Computer and information research scientists | 22% growth; 38,600 jobs | U.S. projection, 2025 to 2035 (BLS) |
Information security analysts | 21% growth; 192,900 jobs | U.S. projection, 2025 to 2035 (BLS) |
Software developers, QA analysts and testers | 10% growth; 1,905,400 jobs | U.S. projection, 2025 to 2035 (BLS) |
LinkedIn reported about 1.3 million "AI-related" opportunities that emerged worldwide between 2023 and 2025. Its list also included about 177,000 AI forensic analysts and 9,000 Head of AI roles. These are LinkedIn's own estimates from its member profiles, which skew toward white-collar and digital work. Many of the largest groups, such as data annotators, are support roles rather than the high-paid research jobs people picture.
The BLS rows come from the 2025 to 2035 projections released August 27, 2026. They describe broad occupations, so they overstate the number of "AI jobs." BLS ties data-scientist growth to demand for developing and using AI. The projection is a forecast, not an observed result.
Several popular titles lack reliable counts. We found no credible 2026 headcount for roles such as MLOps engineer, AI governance specialist or AI security specialist, and prompt engineering shows up in PwC's work as a skill, not as a large standalone occupation. Treat claims that such titles are booming as unproven until a primary source counts them.
AI Salaries in 2026: Pay, Wage Premiums and Role Benchmarks
AI skills are linked to higher advertised pay, but a wage premium is not the same as an average AI salary. A premium compares similar jobs with and without AI skills. A salary benchmark states what a specific occupation earns. Readers need both, and they should not be swapped.
Premiums from three sources
Source | Finding | What it measures |
|---|---|---|
62% average premium (57% a year earlier); 16% in government to 118% in consumer markets | Advertised wages for similar jobs that differ in AI skills; 27 countries, 2025 data | |
28%, or nearly $18,000 per year | U.S. postings with AI skills versus without; 2024 data, published July 2025 | |
AI-exposed occupations' advertised pay up 46% since 2021, versus 25% in the least exposed; premium falls to 2.4 points after adjusting for seniority | U.S. advertised pay growth; a different question from the PwC and Lightcast premiums |
The Indeed numbers come from published summaries because Indeed's page could not be opened during our review, so check the original before quoting them.
The figures differ for good reasons. PwC compares matched profiles across many countries and warns that its estimates do not control for other pay drivers such as education, experience or location. Lightcast compares postings. Indeed measures growth over time for occupations rather than the gap for one skill. A 62% premium does not mean an AI course will raise your pay by 62%.
Official occupational benchmarks (U.S.)
Occupation | Median annual wage | Measure |
|---|---|---|
Data scientists | $120,230 | BLS median annual wage, May 2025 (BLS) |
Software developers | $135,980 | BLS median annual wage, May 2025 (BLS) |
Computer and information research scientists | $140,300 | BLS median annual wage, May 2025 (BLS) |
Information security analysts | $129,180 | BLS median annual wage, May 2025 (BLS) |
These are U.S. wages for actual BLS occupations, in U.S. dollars. They are not "AI engineer salaries," because BLS has no such category. Many people in these occupations do little AI work, and many AI engineers are classified elsewhere.
Context changes the picture. Sector matters: PwC's global premium runs from 16% in government to 118% in consumer markets, and PwC Canada found 94% in consumer markets and 24% in financial services. Seniority matters: Indeed's analysis points to employers paying more for experienced workers in exposed fields. Geography matters: Singapore's sector premiums ranged from 32% to 107%. Advertised pay is also what employers offer, not what people are hired at. Use these numbers to compare sectors and skills, and use your local market data before negotiating.
The AI Skills Employers Want Most in 2026
Employers want three layers of skill: technical AI skills, practical AI literacy, and human skills that AI does not replace. The evidence is strongest for the first and third layers and thinnest for specific tools, so we name only what sources support.
AI engineering and technical skills
Python was the most requested specialized skill in Lightcast's U.S. 2025 data, appearing in 258,674 postings, up about 30% from 2024. Cloud platforms such as Amazon Web Services, scalability and workflow management also ranked high. The World Economic Forum's survey lists AI and big data, networks and cybersecurity, and technological literacy as the fastest-growing skills through 2030 (WEF skills outlook).
We did not find reliable 2026 counts for skills such as retrieval systems, model evaluation or MLOps, so we do not rank them. Agentic AI is the clearest rising technical signal, at 0.23% of U.S. postings.
Applied AI and AI literacy
AI literacy means using AI tools well, checking their output and fitting them into real work. LinkedIn reported 70% year-over-year growth in U.S. jobs requiring AI literacy. It also reported that employees at organizations using LinkedIn Learning were building AI skills 3.4 times faster, which is a vendor-reported finding that may reflect already motivated employers.
Because 51% of AI-skill postings sat outside IT in Lightcast's 2024 data, these skills now appear in marketing, human resources, finance and education. Prompt engineering is one example of an AI skill in PwC's pay analysis, but we found no evidence that it has become a large stand-alone career. Treat it as a useful habit, not a job title to chase.
Complementary human skills
Analytical thinking is the top core skill, cited by 70% of employers in the WEF survey. Resilience, flexibility and agility followed at 67%. PwC found that the most AI-exposed jobs are adding tasks built on human-intensive skills, such as empathy, judgment and creativity, 2.5 times faster than the least exposed jobs.
Why would human skills gain value? When AI handles routine drafting, sorting and calculation, the remaining work leans on deciding what to do, checking whether it is right and dealing with people. WEF also expects 39% of workers' core skills to change by 2030, and PwC found skills in the most AI-exposed jobs changing twice as fast as in the least exposed. Learning how to learn is a skill in itself. For a closer look at tools driving these shifts, see our generative AI statistics guide.
Which Industries Are Hiring the Most AI Talent?
Technology is the most AI-intensive sector in most datasets, but AI hiring is no longer limited to it. In Lightcast's 2024 data, 51% of postings requiring AI skills were outside IT and computer science occupations. In Singapore's 2025 postings, technology, media and telecoms had the largest share of AI hiring at 18.9%, followed by government at 13.5% and financial services at 12.2% (PwC Singapore).
Volume, share, growth and pay concentration are separate questions. We found good, comparable data on pay by sector, shown below.
Sector | AI-skill advertised-wage premium | Context |
|---|---|---|
Consumer markets | 118% | Highest of the sectors PwC analyzed |
Technology, media and telecoms | 84% | Also the largest AI hiring share in Singapore |
Energy, utilities and resources | 75% | |
Manufacturing | 73% | |
Professional services | 67% | |
Financial services | 53% | |
Health industries | 37% | |
Government and public sector | 16% | 107% in Singapore, 24% in the U.S. |
Source: PwC 2026 Global AI Jobs Barometer, 27 countries, 2025 data. The U.S. government figure comes from PwC's U.S. report.
A high premium may mean AI skills are scarce in that sector, not that the sector hires the most AI workers. The government example shows how much geography matters: 16% globally, 24% in the U.S. and 107% in Singapore. We did not find comparable 2026 counts for legal, education or retail hiring, so we make no sector-ranking claim for them.
How Generative AI and AI Agents Are Changing Existing Jobs
Generative AI is changing jobs mainly by changing tasks, not by removing whole occupations. The ILO's modeling concluded that "transformation of jobs is the most likely impact of GenAI" rather than wholesale displacement. The evidence on agents is newer and thinner, but it points the same way.
PwC splits advertised U.S. and global roles into two tracks. In "professionalised" jobs, AI handles the basic work and leaves expert tasks to people. Examples include radiologists and recruiters. In "democratised" jobs, AI takes on complex work so less specialized workers can do more. Examples include IT service managers and medical secretaries. PwC found professionalised roles made up 22% of advertised roles and democratised roles 52%, with 26% showing low AI exposure. Professionalised jobs grew faster (39% versus 17% in postings from 2018 to 2025) and saw faster wage growth (37% versus 26% since 2021).
Agents add a supervision layer. As tools carry out multi-step tasks, people spend more time setting goals, reviewing results and taking responsibility for errors. Employers say they expect this shift: in WEF's survey, about two-thirds plan to hire people with specific AI skills, 85% plan to prioritize upskilling and about 41% plan to reduce staff where AI automates tasks (WEF press release). These are stated plans, not outcomes. For a look at how agents may reshape business software, see our guide to AI agents versus SaaS.
In software, the evidence is mixed. Indeed reported in July 2026 that U.S. software-development postings were up about 15% since late February 2025, while all postings were down about 7%, and that 71% of the May 2025 to May 2026 increase came from senior roles. Indeed noted that correlation does not prove causation. BLS projects growth for software developers but expects productivity gains from AI to dampen demand in some occupations such as office support and sales.
Entry-Level Jobs and the Changing Career Ladder
Evidence that AI is hurting entry-level work is real but narrow, and it does not show that AI has killed entry-level jobs. The strongest finding applies to young workers in specific exposed occupations, while other data shows entry-level jobs growing when they demand higher-level skills.
Case study 1: Stanford's payroll study
Researchers Erik Brynjolfsson, Bharat Chandar and Ruyu Chen used payroll records from ADP. They found a 13% relative employment decline for workers aged 22 to 25 in the most AI-exposed occupations since generative AI spread. Older workers in the same occupations had stable or growing employment. The declines were concentrated where AI automates work rather than assists it, and adjustment came through jobs, not pay. The authors call the evidence "early" and consistent with an AI effect, not proof of one (Stanford working paper, August 2025).
Case study 2: PwC's 2.4 million entry-level jobs
PwC analyzed 2.4 million entry-level U.S. jobs. Among the most AI-exposed quartile, "seniorised" entry-level roles grew 35% from 2019 to 2025, while other entry-level roles fell 10%. AI-exposed junior roles were seven times more likely to demand traditionally senior skills such as leadership and strategic thinking (PwC global findings). The ladder is not gone. The first rung is higher.
Case study 3: Indeed's software rebound
Indeed found that from May 2022 to May 2026, higher AI exposure went with steeper posting declines. From May 2025 to May 2026, the pattern reversed and AI-exposed occupations rebounded faster. Because 71% of the rise in software-development postings came from senior roles, most of the recovery was not in junior hiring (Indeed Hiring Lab, July 8, 2026).
What this means
Broader forces also hit young workers. The OECD's Employment Outlook 2026, updated July 31, 2026, said that employment is resilient but showing signs of weakening, and that young people face particularly difficult entry conditions (OECD summary). The routine tasks that once taught beginners, such as drafting, basic coding and data cleanup, are the tasks AI performs best. That may remove apprenticeship steps even where headcount holds. The same tools can also speed up learning if employers design for it.
Which Jobs Are Most Exposed to AI — and Does Exposure Mean Replacement?
Exposure does not mean replacement. Exposure measures how many of a job's tasks AI could perform. Whether that leads to help, fewer hours or job loss depends on costs, rules, demand and how employers redesign work.
Term | Meaning | Evidence example |
|---|---|---|
Exposure | AI could perform some tasks | ILO: about 1 in 4 workers globally are in jobs with some generative-AI exposure; 3.3% in the highest tier |
Augmentation | AI helps a worker do the job | PwC "professionalised" jobs such as radiologists and recruiters |
Partial task automation | AI does some tasks, people do the rest | BLS expects AI productivity gains to dampen demand in office support, sales, arts and design |
Substitution | AI does work that people were hired for | Stanford ADP study: declines for ages 22 to 25 where AI automates tasks |
Full occupational displacement | An occupation largely disappears | Not documented as AI-caused in the sources reviewed here |
Clerical work has the highest exposure in the ILO's index. Women face higher exposure in the top tier: 4.7% of women's employment versus 2.4% of men's globally, and 9.6% versus 3.5% in high-income countries. Wealth matters as well. Some generative-AI exposure covers 34% of employment in high-income countries and 11% in low-income countries, in part because richer economies have more digital, office-based work (ILO, May 20, 2025).
Myths versus facts
Myth: High exposure means the job will disappear. Fact: ILO modeling says transformation is the most likely effect.
Myth: AI jobs are only for engineers. Fact: 51% of AI-skill postings were outside IT in 2024 data.
Myth: AI has already wiped out entry-level work. Fact: Evidence shows pressure on some young workers, rising skill demands for others, and several non-AI causes.
AI Workforce Impact: Job Creation, Displacement, Productivity and Inequality
The best-supported conclusion is that AI is reshaping jobs faster than it is eliminating them, with the risks falling unevenly. Observed evidence shows productivity and pay differences, targeted pressure on entry-level work and a widening gap between workers who gain AI skills and those who do not.
Job creation and displacement. WEF projects 170 million jobs created and 92 million displaced by 2030, a net gain of 78 million. These figures cover all structural trends, including demographics, geopolitics, the green transition and technology, not AI alone. WEF's AI-specific estimate is much smaller: AI and information-processing technologies are expected to create 11 million jobs and displace 9 million. Robotics and autonomous systems are expected to be the largest net displacer, with a net decline of 5 million jobs. These are employer-survey forecasts.
Productivity. PwC found productivity growth was 40% higher in the most AI-exposed companies than the least exposed. Among sectors, productivity grew 34% from 2018 to 2025 in the most exposed versus 24% in the least. The most exposed companies also grew headcount by 52% versus 36% and wages by 24% versus 17%. These are associations. Strong firms may adopt AI first, so AI may not be the only cause.
Inequality. PwC calls the result a "two-track" labor market. The ILO shows higher top-tier exposure for women and a large gap between rich and poor countries in exposure, which is not the same as readiness. Workers in lower-income economies may face less near-term exposure but also fewer chances to benefit. Education, age and local access to training all affect who gains, but we found little comparable 2026 data on these links, so we avoid firm claims.
What the evidence supports | What it does not support |
|---|---|
AI skills carry higher advertised pay | That AI skills guarantee higher earnings |
Productivity is higher at AI-exposed firms | That AI alone caused the gain |
Some early-career pressure in exposed jobs | That AI has ended entry-level hiring |
Net job growth in WEF's AI-specific forecast | That forecasts are certain |
Avoid false precision. Every number above is a rough guide to direction and size.
What AI Jobs Statistics Mean for Job Seekers and Career Changers
Learning AI is worthwhile for most workers, but becoming an AI engineer is right for only some. The statistics show broad demand for AI literacy, a premium for AI skills in advertised pay and rising skill requirements even in entry-level roles. Use this six-step framework:
Build baseline AI literacy. Learn to use AI tools, check their output and know their limits. Demand for AI literacy grew 70% in the U.S. in LinkedIn's data.
Pair AI with domain expertise. With 51% of AI-skill postings outside IT, a finance, health, marketing or legal specialist who uses AI well fits real demand.
Learn workflow redesign. Show you can take a messy process and make it faster, with a person still in charge.
Strengthen judgment and evaluation. Analytical thinking is the top employer-rated skill in WEF's survey.
Add technical depth only if your path needs it. Python, data handling and cloud skills lead technical postings. BLS lists a bachelor's degree as typical for data scientists and a master's for research scientists, but a degree is a norm, not a law.
Build proof of work. Keep examples, results and descriptions of how you checked quality.
You do not need to code for every AI-related role, but you do need it for building and deploying models. Use premiums to compare options, not to forecast your own pay. If you are starting from education, see our guide to AI in education. The signals to watch are the share of postings in your field that list AI skills, entry-level postings in your target role and pay ranges in your region.
What AI Jobs Statistics Mean for Employers and Workforce Leaders
Most organizations need AI-capable employees more than they need many AI engineers. The data shows skill demand spreading across functions, skills changing about twice as fast in exposed jobs, and an advertised-pay premium that affects hiring competition.
Hire specialists when you build or run models, secure systems or manage data pipelines. Competition is real: BLS projects about 35% data-scientist growth through 2035.
Upskill existing staff for everyday AI use. In WEF's survey, 85% of employers prioritize upskilling.
Redesign roles where tasks are routine and checkable, and define who reviews AI output.
Review pay by role and market. The PwC premium is a signal of scarcity, not a pay formula, and it does not control for experience or location.
Protect the entry-level pipeline. If AI takes routine tasks, create structured ways for juniors to learn judgment, such as shadowing reviews. PwC's "seniorised" findings show junior roles demanding more.
Govern before you automate. Keep human sign-off for decisions affecting people, money or safety.
Use outside experts for short, specialized work, and build internal skills for continuing work.
For benchmarks on company-level adoption, see our enterprise AI statistics.
AI Jobs Outlook Through 2030–2035
The outlook is a mix of well-measured trends and uncertain forecasts. Separating them helps you decide what to act on now.
What we know now
AI skills are in a growing, still-minority share of U.S. postings (2.5% to 2.8% in 2025).
Advertised pay is higher for AI skills in several datasets, with large variation by sector and country.
Skills are changing faster in exposed jobs.
Early-career pressure exists in some exposed occupations, but its cause is not settled.
What remains a forecast
BLS projects U.S. employment to grow 3.5% from 2025 to 2035, from 170.3 million to 176.2 million jobs, with data scientists among the fastest-growing occupations at about 35%.
WEF's 2030 estimates: 11 million AI-related jobs created, 9 million displaced, and 78 million net new jobs from all trends.
Any claim about how many jobs agents will replace.
Indicators to monitor
AI-skill share of postings (Lightcast, PwC and Indeed series)
AI pay premiums, adjusted for seniority
Entry-level hiring in exposed occupations
Agentic AI demand beyond 0.23% of postings
Productivity and headcount at AI-exposed firms
Adoption outside the technology sector
Degree and experience requirements
FAQ
How many AI jobs are there in 2026?
No one knows exactly, because there is no official definition or worldwide count. In U.S. data for 2025, PwC counted more than 1.12 million postings requiring AI skills. Postings are job ads, not people employed.
Is demand for AI jobs increasing in 2026?
Yes, by most measures. U.S. AI-skill postings rose 55% to 66% in 2025, and Indeed's AI-mention share roughly doubled in the year to August 2026. Demand is growing from a small base.
What percentage of job postings require AI skills?
In the U.S. in 2025, 2.5% (Lightcast) to 2.8% (PwC) of postings listed or required AI skills. Indeed's broader keyword measure showed 6.74% mentioning AI on August 31, 2026.
Which AI jobs are in highest demand?
Fast-growing roles include AI engineers, forward-deployed engineers and data annotators in LinkedIn data, and AI/machine-learning specialists in WEF's survey. BLS projects about 35% growth for U.S. data scientists through 2035.
What do AI-related jobs pay?
It depends on the occupation. BLS median annual wages in May 2025 were $120,230 for data scientists, $135,980 for software developers and $140,300 for computer and information research scientists. These are not AI-specific salaries.
Do AI skills increase salary?
Advertised pay is higher for AI skills in several datasets: a 62% average premium in PwC's data and 28% in Lightcast's U.S. postings. Neither guarantees your pay will rise.
Which AI skills are most valuable in 2026?
Python leads technical demand in U.S. postings, and AI literacy is spreading quickly. Employers also rate analytical thinking and resilience highly. We found no reliable ranking of narrower skills.
Do you need a degree to get an AI job?
Not always. BLS lists a bachelor's degree as typical for data scientists and a master's for computer and information research scientists. Many AI-enabled roles value skills and proof of work, but rules vary by employer.
Do you need coding skills to work with AI?
You need coding to build and deploy models. You do not need it to use AI well in many roles, since 51% of AI-skill postings were outside IT in 2024 data.
Is prompt engineering still valuable?
As a skill, yes. PwC includes it among AI skills with a pay premium. As a stand-alone career, we found no evidence of large demand.
Is AI replacing entry-level jobs?
Not across the board. A Stanford study found a 13% relative employment decline for ages 22 to 25 in highly exposed occupations, while PwC found AI-exposed entry-level roles growing but demanding senior-level skills.
Which jobs are most exposed to AI?
The ILO finds clerical occupations most exposed to generative AI. Exposure means tasks could be done by AI, not that the job will vanish.
Will AI create more jobs than it eliminates?
WEF expects 11 million jobs created and 9 million displaced from AI and information processing by 2030, a net gain of 2 million. This is a forecast. WEF's larger 170 million and 92 million figures cover all trends.
Which industries hire the most AI talent?
Technology leads in most datasets, but half of AI-skill postings were outside IT in 2024 data. In Singapore, technology, media and telecoms, government and financial services had the largest AI hiring shares.
What should workers learn to stay employable as AI adoption grows?
Learn AI tools and their limits, pair them with domain knowledge, practice judgment and evaluation, and keep proof of your work.
Key Takeaways
AI jobs statistics depend on definitions: postings, mentions, exposure and displacement are different measures.
AI-skill demand is rising fast but remains about 2.5% to 2.8% of U.S. postings in strict measures.
Advertised AI wage premiums are large but do not equal average AI salaries.
AI skills are spreading beyond tech, with 51% of AI-skill postings outside IT in 2024 data.
Entry-level evidence is mixed: pressure on some young workers, and rising skill requirements for others.
WEF's 170 million and 92 million job figures cover all trends, not AI alone.
Exposure is not replacement, and productivity gains are associations, not proof of cause.
Both job seekers and employers benefit from combining AI literacy, domain expertise and judgment.
Actionable Next Steps
Check which of your tasks AI could assist, and try one tool on one real task this week.
Look up your occupation in the BLS Occupational Outlook Handbook for pay and outlook.
Add one visible AI-assisted project to your portfolio, with notes on how you checked quality.
Track AI-skill share in postings for your target role over three months.
If you manage a team, list roles by routine tasks and decide which to redesign first.
Set a review rule for any AI output that affects people, money or safety.
Glossary
AI job: A loosely used term for a role that builds or relies on AI. It has no standard definition.
AI literacy: The ability to use AI tools effectively and judge their output.
Generative AI: AI that creates text, images, code or other content.
Agentic AI: AI that plans and takes multi-step actions toward a goal with limited supervision.
AI agent: A software system that performs tasks on a user's behalf using AI.
Machine learning: Methods in which software learns patterns from data.
LLM (large language model): A model trained on large amounts of text to produce language.
MLOps: Practices for deploying and maintaining machine-learning models.
Augmentation: Using AI to help a person do a task better.
Automation: Having technology perform a task without a person.
Occupational exposure: The share of an occupation's tasks that AI could perform.
AI skill premium: The extra advertised pay linked to AI skills, relative to similar jobs without them.
Reskilling: Training someone for a different role.
Upskilling: Building new skills within a current role.
Skills-based hiring: Hiring based on demonstrated skills more than degrees.
Sources & References
PwC. 2026 Global AI Jobs Barometer (press release). June 15, 2026.
PwC. 2026 Global AI Jobs Barometer: Global Findings. 2026.
PwC. 2026 AI Jobs Barometer: United States. 2026.
PwC. 2026 AI Jobs Barometer: Canada. 2026.
PwC Australia. PwC's 2026 AI Jobs Barometer. June 18, 2026.
PwC Singapore. AI roles command wage premiums across sectors in Singapore. June 15, 2026.
Lightcast. Four Takeaways from the 2026 Stanford AI Index. April 2026.
Lightcast. AI Skills Command 28% Salary Premium as Demand Shifts Beyond Tech Industry. July 23, 2025.
LinkedIn Economic Graph. Building a Future of Work That Works (Labor Market Report). January 2026 (copy hosted by ADAPT).
Indeed Hiring Lab. AI Tracker data. Data through August 31, 2026.
Indeed Hiring Lab. AI and Job Postings: From Destruction to Creation?. July 8, 2026.
Indeed Hiring Lab. AI Exposure Isn't Squeezing Advertised Pay in the US — It's Boosting It. September 17, 2026.
U.S. Bureau of Labor Statistics. Employment Projections: 2025–2035. August 27, 2026.
U.S. Bureau of Labor Statistics. Occupational Outlook Handbook: Data Scientists. Updated August 27, 2026.
U.S. Bureau of Labor Statistics. Occupational Outlook Handbook: Software Developers, Quality Assurance Analysts, and Testers. 2026.
U.S. Bureau of Labor Statistics. Occupational Outlook Handbook: Computer and Information Research Scientists. 2026.
U.S. Bureau of Labor Statistics. Occupational Outlook Handbook: Information Security Analysts. 2026.
World Economic Forum. The Future of Jobs Report 2025: Digest. January 7, 2025.
World Economic Forum. Future of Jobs Report 2025: Jobs Outlook. January 2025.
World Economic Forum. Future of Jobs Report 2025: Skills Outlook. January 2025.
World Economic Forum. Future of Jobs Report 2025: 78 million new job opportunities by 2030. January 2025.
International Labour Organization. Generative AI and Jobs: A Refined Global Index of Occupational Exposure. May 20, 2025.
Brynjolfsson, E., Chandar, B., Chen, R. Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence. Stanford Digital Economy Lab / SIEPR. August 2025.
OECD. OECD Employment Outlook 2026 (EU Digital Skills and Jobs Platform summary). Updated July 31, 2026.
