No official statistic found for this article counts jobs created by AI, so the claim that AI is a catalyst for job opportunities rests on three weaker kinds of evidence: economic history, employer forecasts, and early payroll data. The World Economic Forum's Future of Jobs Report 2025 projects 170 million jobs created and 92 million displaced worldwide by 2030, a net gain of 78 million, but that figure covers all forces reshaping work, not AI alone. A Stanford payroll study through June 2026 adds that hiring of young workers in exposed jobs has lagged.
Key takeaways
- Most current U.S. employment is in job specialties that did not exist in 1940, according to an economic-history study published in The Quarterly Journal of Economics in 2024.
- The World Economic Forum's forecast of 78 million net new jobs by 2030 comes from a survey of more than 1,000 employers and is not specific to AI.
- Technology roles lead the forecast's fastest-growing list, while clerical roles are expected to see the largest absolute decline.
- Early payroll data show young workers' hiring lagging in AI-exposed occupations, which the authors treat as descriptive and not proof of cause.
- A headline about jobs created needs a time frame, a definition of "job," and a count of jobs lost alongside it.
Three kinds of evidence, and what each can show
Statistics agencies publish employment by occupation, and the ones reviewed here have no category for "AI-created jobs." That leaves researchers to look from three angles. History shows how new work has appeared after earlier technologies. Forecasts record what employers expect. Payroll data show what has already changed. The table compares them; each has a limit that the others do not share.
| Evidence | Source | Reported finding | Main limit |
|---|---|---|---|
| History, 1940 to 2018 | Autor, Chin, Salomons, and Seegmiller (NBER and QJE) | Most current employment is in specialties introduced after 1940 | Covers past technologies, not generative AI |
| Employer forecast, 2025 to 2030 | World Economic Forum | 170 million created, 92 million displaced, 78 million net | Opinion of surveyed employers; all drivers, not only AI |
| Payroll records through June 2026 | Brynjolfsson, Chandar, and Chen | Hiring gap for young workers in exposed occupations | Descriptive; one payroll provider's sample |
What history says about new work
David Autor, Caroline Chin, Anna Salomons, and Bryan Seegmiller studied how new job specialties appeared in U.S. employment between 1940 and 2018. Their working paper, later published in The Quarterly Journal of Economics, finds that the majority of current employment is in specialties introduced after 1940. The source of new work has shifted: from middle-paid production and clerical jobs between 1940 and 1980 to high-paid professional jobs and, to a lesser degree, low-paid service jobs since 1980.
The authors separate two kinds of innovation. Those that augment an occupation's output raise demand for its labor, and those that automate its tasks lower it. Both kinds tend to appear in the same occupations, and the authors report that the demand-reducing effect of automation has strengthened over four decades while the demand-raising effect of augmentation has not. History therefore supports the idea that new work appears, and it does not support the idea that the new work arrives in proportion to what is lost.
What employers expect by 2030
The World Economic Forum's Future of Jobs Report 2025, published in January 2025, draws on more than 1,000 employers representing over 14 million workers in 22 industry clusters and 55 economies. It estimates that job creation and destruction from structural change will equal 22 percent of today's jobs: new jobs equal to 14 percent of current employment (170 million), displacement of 8 percent (92 million), and net growth of 7 percent (78 million).
Within that, the report lists Big Data Specialists, FinTech Engineers, AI and Machine Learning Specialists, and Software and Applications Developers among the fastest-growing roles in percentage terms. Frontline roles such as farmworkers and delivery drivers add the most jobs in absolute volume. Clerical and secretarial workers, including cashiers and administrative assistants, are expected to shrink the most. The report's drivers include aging, the green transition, and cost of living as well as technology, so reading the net figure as an AI figure overstates what it measures. It is also a survey of expectations, not a count of jobs.
What payroll data show so far
Payroll data are the only one of the three sources that records what has already happened. The Stanford Digital Economy Lab study, revised August 12, 2026, finds no widespread job displacement but a hiring shortfall for workers aged 22 to 25 in AI-exposed occupations, and its authors treat the result as an early indicator, not a causal estimate. That is the one place where the evidence points to narrowing opportunity, at entry level, while forecasts show new roles elsewhere.
Reading a claim that AI creates jobs
- Does the number count jobs created, or net of jobs lost, and over what years?
- Is it measured from payroll or census records, or is it an expectation from a survey?
- Does it isolate AI, or does it bundle AI with other drivers such as aging or the energy transition?
- Which workers gain the new roles, and which workers lose the old ones?
- Does "job" mean a new occupation, or an existing role with changed tasks?
This article reports published research and projections and does not forecast employment for any occupation or person.
FAQ
Does AI create more jobs than it removes?
The forecast with the most coverage says yes in net terms by 2030, but it measures all drivers of change, and it reflects employer expectations. Measured data through mid-2026 show no widespread displacement, though entry-level hiring has lagged in exposed occupations.
Which jobs does the forecast expect to grow fastest?
The World Economic Forum report lists technology roles such as Big Data Specialists and AI and Machine Learning Specialists at the top in percentage terms. Care, education, and frontline roles add large absolute numbers.
Are there Canadian figures for AI-created jobs?
None turned up in the sources reviewed. Canadian statistics describe employment by occupation and exposure, not jobs created by AI, and the World Economic Forum figures are global.
What is the difference between a new job and a changed job?
A new job is an occupation or specialty that did not exist before, as in the 1940 to 2018 study. A changed job keeps its title while its tasks shift, and exposure studies count that separately.
Related reading
- AI and employment: challenges and opportunities
- Whether AI can replace human workers
- The question of widespread unemployment
- The future job market and AI
More in AI and Society.
Last verified: October 2026
Sources
- New Frontiers: The Origins and Content of New Work, 1940-2018, Autor D, Chin C, Salomons A, Seegmiller B, NBER Working Paper 30389 and The Quarterly Journal of Economics 139(3), 2024
- Future of Jobs Report 2025, World Economic Forum, January 2025
- Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence, Brynjolfsson E, Chandar B, Chen R, Stanford Digital Economy Lab, revised August 12, 2026