Will AI create more jobs than it destroys?

The most authoritative and recent large-scale projection comes from the World Economic Forum’s Future of Jobs Report 2025 (released early 2025, covering 2025–2030). It estimates:

92 million jobs displaced globally due to AI, automation, and related macro trends.

170 million new jobs created in the same period.

Net gain of 78 million jobs (roughly +7% of current global employment).

This marks a shift from their 2023 report (which leaned more pessimistic) and aligns with patterns seen in previous technological shifts (e.g., computers, internet, ATMs → more bank tellers over time due to productivity gains and new services).

Other major 2025–2026 analyses largely point in a similar direction, though with caveats about the transition:

Goldman Sachs (updated 2025 analyses): AI could displace 1–4 million jobs per year in coming years (potentially reducing job growth temporarily), but ultimately create more through productivity effects, new industries, and expanded output. They project only modest, transitory unemployment rises (≈0.5 percentage points) during adjustment, with AI-exposed occupations actually showing faster wage and job growth in many cases.

PwC’s 2025 Global AI Jobs Barometer: Industries most exposed to AI see 3× higher revenue-per-employee growth and wages rising 2× faster than less-exposed sectors. Skills in AI-exposed jobs change 66% faster, but workers become more valuable overall—not less.

McKinsey and others: While 30–60% of tasks (or up to 57% of U.S. work hours) could be automated, full job elimination is rarer; most roles get augmented (AI handles rote parts → humans focus on judgment, creativity, relationships). They emphasize that AI boosts lower-skilled workers’ productivity more in many cases, potentially narrowing gaps.

Short-Term Reality (2025–2026)

Displacement is already visible and painful in pockets:

Tech layoffs in 2025 directly tied to AI reached tens of thousands (e.g., ~78,000 reported in first half of 2025).

Entry-level white-collar roles (data entry, basic analysis, junior coding, some customer service) face the steepest hits—some CEOs (Anthropic, Ford, etc.) warn of 10–50% cuts in certain categories.

Anticipatory layoffs (companies cutting staff expecting future AI efficiency) outpace actual performance-based ones so far.

Yet overall unemployment remains low in most advanced economies, and total job growth continues (U.S. BLS projects modest but positive net additions through 2035–2036).

Long-Term Outlook (2030+)

History shows technology tends to create more net jobs than it destroys once productivity gains feed into economic expansion, new industries, and demand for non-automatable human work (empathy, complex problem-solving, physical dexterity in unpredictable settings, leadership). AI looks likely to follow this pattern—potentially accelerating GDP growth (Goldman Sachs: up to +7% globally) and spawning roles we can’t yet name (AI system designers, ethics auditors, human-AI collaboration specialists, advanced creative oversight, etc.).

The big if: Whether net creation outpaces destruction depends on:

Speed of adoption (faster = sharper short-term pain).

Reskilling scale (governments, companies, individuals investing in training).

Policy choices (retraining programs, income support during transition, incentives for human-complementary AI use).

Geographic/sector inequality (advanced economies face 60%+ task exposure; emerging markets lower but growing fast).

Bottom Line

On current evidence and projections from 2025–2026 reports (WEF, Goldman, PwC, etc.), yes—AI is more likely to create more jobs than it destroys in aggregate over the next 5–10 years, with a projected global net positive by 2030. But the transition will be disruptive: millions will face displacement (especially routine cognitive/white-collar tasks), entry barriers may rise temporarily for new workers, and success hinges on proactive upskilling and adaptation.

The real risk isn’t mass permanent unemployment—it’s a bumpy, unequal adjustment period where those who learn to work with AI thrive, while those who don’t get left behind. If history (and the data) are any guide, the economy will ultimately generate new work we need humans for, just as it always has. The question is how well society manages the ride there.