Will AI create or take away more job opportunities?

Leans yesSays it is contested

[1][2][3][4]AI is projected to create more job opportunities than it eliminates over the longer term (e.g., through 2030), but it will cause substantial disruption, job churn, and uneven effects in the short-to-medium term. The net outcome depends heavily on how societies, companies, and workers adapt—through reskilling, augmentation (using AI to enhance human work) rather than pure replacement, policy support, and the creation of new roles and industries.[1]

Key Projections and Expert Views

Major analyses point to net job gains, alongside large-scale shifts:

•The World Economic Forum’s Future of Jobs Report 2025 (based on surveys of ~1,000 large employers representing 14 million workers) projects that technological and other trends, with AI as a major driver, will create about 170 million new jobs globally by 2030 while displacing 92 million, for a net gain of 78 million. AI ranks as a top transformer of business (86% of companies expect it to reshape operations). Growing roles include AI/ML specialists, big data specialists, fintech engineers, and green/energy transition jobs. Declining ones include certain clerical/admin roles (e.g., postal clerks, secretaries, payroll), and some creative/admin positions newly pressured by generative AI (e.g., graphic designers, legal secretaries in some analyses). Many employers plan to hire for AI skills while reducing headcount in automatable areas, with strong emphasis on retraining.[1]

•Gartner research indicates AI will create more jobs than it eliminates beginning in 2028, though it will disrupt career paths and require rethinking talent pipelines (e.g., fewer traditional entry-level rungs as AI handles routine tasks).[3]

•Goldman Sachs and others see meaningful displacement potential (e.g., tens of millions of workers exposed or shifted over a decade in the US), but view labor market headwinds as temporary because AI boosts productivity, lowers costs, raises demand, and creates new roles—especially where it augments rather than fully substitutes labor. Early data showed a modest net drag (e.g., reduced monthly payroll growth), partially offset by gains in augmentation-heavy areas. Effects have hit younger/less-experienced workers harder so far.[2]

•Other views (e.g., BCG, ILO, various economists) emphasize that a large share of jobs (often 50%+) will be reshaped (tasks change) rather than fully eliminated. Exposure ≠ elimination—many roles retain human elements like judgment, empathy, physical presence, or complex oversight. Estimates of pure replacement are lower (e.g., 10-15% of US jobs over several years in some BCG analysis).[5]

Economists like David Autor (MIT) note that technology automates, complements, and creates new work/expertise. Outcomes hinge on implementation: pure automation can destroy jobs, while tools enabling broader expertise or new demand (e.g., in healthcare) can expand them. Historical patterns support net creation long-term, though recent decades showed stronger automation relative to new-task creation in some US data.[6]

Current Evidence (as of mid-to-late 2020s)

Mass unemployment from AI has not materialized. Overall labor markets remain resilient in many places, with AI contributing to new hiring even as it pressures specific areas:

•New AI-related roles have grown substantially: LinkedIn data and related analyses cite ~1.3 million new AI-enabled/related jobs (e.g., AI engineers, data annotators, forward-deployed engineers) plus hundreds of thousands in data centers and infrastructure. Professional occupations tied to AI development (engineers, data scientists, etc.) have expanded. Some estimates put AI-created US jobs around 1 million, outpacing attributed layoffs (~200,000 in certain tallies).[7]

•Displacement is real but concentrated: AI-cited layoffs have risen (tens of thousands annually in the US via trackers like Challenger, Gray & Christmas, sometimes 20-40% of monthly cuts in peaks), hitting tech, admin, customer service, content/creation, and entry-level white-collar roles. Stanford/ADP payroll analyses show reduced opportunities especially for young workers (ages 22-25) in highly exposed occupations—often via slower hiring rather than mass firings of incumbents. Experienced workers are more insulated. Some companies that cut for AI have later rehired or regretted pure replacement due to quality/judgment gaps.[4]

•Broader patterns: Employment in many AI-exposed areas has still grown overall (or declined less than feared). AI often reduces hiring needs for routine tasks without proportional layoffs, consolidates roles, and boosts productivity/wages for those who use it effectively (wage premiums for AI skills reported). Adoption remains uneven across sectors.[8]

Historical Context

Past technologies (steam, electricity, computers, ATMs, industrial robots) caused short-term displacement in specific roles/sectors (e.g., handloom weavers, elevator operators, certain manufacturing) but ultimately expanded total employment through productivity gains, lower costs, new industries, and complementary jobs. Roughly 60% of current US jobs involve types of work that did not exist in 1940. New work has often favored younger, skilled workers. Agriculture and manufacturing employment shares plummeted while overall jobs and living standards rose. AI follows a similar dual path (substitution + demand creation), though its speed, cognitive scope, and potential for broad task automation raise unique questions about pace and inequality.[9]

Caveats and Challenges

•Uneven impacts: Routine cognitive/admin tasks, junior roles, and certain creative/service jobs face higher risk. Physical, interpersonal, highly skilled judgment, trades, healthcare (with human elements), and AI-orchestration roles are more resilient or growing. Developing economies or outsourcing-dependent ones may face different pressures. Inequality could rise without intervention if gains concentrate among high-skill/AI-fluent workers and capital owners.

•Transition frictions: Skills mismatch is major—displaced workers may not easily move into new roles. Entry-level pipelines are narrowing in some fields. Unemployment effects are expected to be temporary and modest relative to normal churn in optimistic scenarios, but painful for affected individuals/groups.

•Uncertainty: Faster capability jumps (toward more agentic or general AI), slower adaptation, or heavy replacement-focused deployment could worsen short-term losses. Productivity booms could also create unexpected demand. Long-term (beyond 2030) depends on AGI trajectories, which remain speculative.

•Company behavior: Some pursue cost-cutting via automation; others focus on augmentation and growth. Evidence suggests augmentation often yields better returns.

Bottom Line and Implications

Net creation is the more likely long-run outcome, consistent with history and leading forecasts, but “more jobs” does not mean “the same jobs” or a painless transition. Expect significant churn (jobs destroyed + created + transformed), rising demand for AI literacy + uniquely human skills (creativity, critical thinking, emotional intelligence, adaptability, domain expertise + AI collaboration), and infrastructure/support roles.

What helps maximize creation and minimize harm:

•Individuals: Upskill in AI tools, complementary skills, and lifelong learning. Focus on areas hard to fully automate.

•Companies: Prioritize augmentation, internal mobility/reskilling, and redesigning work around human-AI teams rather than pure headcount cuts.

•Policymakers/educators: Invest in education/training systems, portable benefits, safety nets for transitions, incentives for job-creating AI applications, and monitoring labor markets. Address infrastructure (data centers, energy) which itself creates jobs.

In short, AI is a powerful general-purpose technology that expands the economic pie and enables new forms of work, much like predecessors—but only if humans steer it toward complementarity and broad opportunity rather than pure displacement. The “will it create or take away more?” answer leans create (net), with the critical variable being adaptation speed and quality.

Citations (9)

1arstechnica.comNews and mediaquality 0.95
https://arstechnica.com/ai/2025/01/ai-could-create-78-million-more-jobs-than-it-eliminates-by-2030-report/
Backs: “[[1]](https://arstechnica.com/ai/2025/01/ai-could-create-78-million-more-jobs-than-it-eliminates-by-2030-report/)”
statistic, article, points to yes. Claim: WEF projects 170 million jobs created and 92 million displaced by 2030, a net gain of 78 million, with AI a major driver.
2goldmansachs.comCompanies
https://www.goldmansachs.com/insights/articles/the-jobs-ai-is-likely-to-boost-and-those-it-may-disrupt
Backs: “[[2]](https://www.goldmansachs.com/insights/articles/the-jobs-ai-is-likely-to-boost-and-those-it-may-disrupt)”
evidence, article, neutral. Claim: Goldman Sachs sees meaningful displacement but views headwinds as temporary; early data show a modest net drag, offset by augmentation gains.
3gartner.comCompanies
https://www.gartner.com/en/newsroom/press-releases/2026-05-13-gartner-hr-research-reveals-ai-will-create-more-jobs-than-it-eliminates-beginning-in-2028
Backs: “[[3]](https://www.gartner.com/en/newsroom/press-releases/2026-05-13-gartner-hr-research-reveals-ai-will-create-more-jobs-than-it-eliminates-beginning-in-2028)”
attribution, official document, points to yes. Claim: Gartner says AI will create more jobs than it eliminates beginning in 2028, while disrupting career paths.
4theguardian.comNews and mediaquality 0.75
https://www.theguardian.com/technology/2026/aug/12/ai-job-destruction
Backs: “[[4]](https://www.theguardian.com/technology/2026/aug/12/ai-job-destruction)”
evidence, article, points to no. Claim: AI-cited layoffs are rising and hit young and entry-level workers; some companies later rehired after cutting for AI.
5cbsnews.comNews and mediaquality 0.88
https://www.cbsnews.com/news/artificial-intelligence-ai-jobs-boston-consulting-group/
Backs: “Effects have hit younger/less-experienced workers harder so far. Other views (e.g., BCG, ILO, various economists) emphasize that a large share of jobs (often 50%+) will be reshaped (tasks change) rather than fully eliminated.”
statistic, article, neutral. Claim: BCG analysis suggests pure replacement is lower (10-15% of US jobs), with many more jobs reshaped rather than eliminated.
6mit.eduAcademic
https://news.mit.edu/2026/technology-creates-jobs-young-skilled-workers-ai-0521
Backs: “Estimates of pure replacement are lower (e.g., 10-15% of US jobs over several years in some BCG analysis). Economists like David Autor (MIT) note that technology automates, complements, and creates new work/expertise.”
attribution, article, neutral. Claim: David Autor notes technology automates, complements and creates work; outcomes depend on implementation, and recent decades favored automation.
7weforum.orgNonprofits and think tanks
https://www.weforum.org/stories/jobs-and-the-future-of-work/ai-has-already-added-1-3-million-new-jobs-according-to-linkedin-data/
Backs: “New AI-related roles have grown substantially: LinkedIn data and related analyses cite ~1.3 million new AI-enabled/related jobs (e.g., AI engineers, data annotators, forward-deployed engineers) plus hundreds of thousands in data centers and infrastructure.”
statistic, article, points to yes. Claim: LinkedIn data cite about 1.3 million new AI-related jobs, outpacing attributed layoffs.
8rand.orgNonprofits and think tanks
https://www.rand.org/pubs/commentary/2025/10/ai-is-making-jobs-not-taking-them.html
Backs: “Some companies that cut for AI have later rehired or regretted pure replacement due”
evidence, article, points to yes. Claim: Employment in many AI-exposed areas has still grown; AI reduces hiring needs for routine tasks without proportional layoffs and boosts wages for AI users.
9improvado.ioCompanies
https://improvado.io/blog/ai-job-loss-employment-impact
Backs: “Past technologies (steam, electricity, computers, ATMs, industrial robots) caused short-term displacement in specific roles/sectors (e.g., handloom weavers, elevator operators, certain manufacturing) but ultimately expanded total employment through productivity gains, lower costs, new industries, and complementary…”
background, article, points to yes. Claim: Past technologies expanded total employment long-term; about 60% of current US jobs involve work types that did not exist in 1940.