"AI will destroy 300 million jobs." You know that headline. What you read less often: US layoff statistics so far show about 173,000 job cuts attributed to AI, while the market pays a 62% wage premium for AI skills.
Between doomsday forecasts and denial lies a body of data that's far more interesting than either. For this article, I compiled the big forecasts (WEF, Goldman Sachs), the hard actuals (Challenger, Stanford, company cases), Anthropic's usage data, and the German surveys, and had everything cross-checked.
- Forecast vs. reality: the WEF expects 170M new and 92M displaced jobs worldwide by 2030 (net +78M). In actual US data, roughly 173,600 job cuts have been attributed to AI since 2023
- The disruption hits juniors: 22- to 25-year-olds in AI-exposed occupations lost 13% employment, young developers nearly 20%, while experienced workers gained (Stanford). Entry-level roles at tech majors: -65% since 2019
- The upside is real: a 62% wage premium for AI skills (PwC, 2026), 1.3M new AI jobs per LinkedIn, and AI-exposed companies are growing headcount faster than non-exposed ones
1. The Key Numbers at a Glance
2. The Big Forecast: 170 Million New, 92 Million Displaced Jobs
The most-quoted labor market forecast comes from the World Economic Forum's Future of Jobs Report, based on a survey of over 1,000 employers representing 14 million workers across 55 economies:
The bottom line: the WEF expects a net gain of 78 million jobs by 2030. The real number behind the number is the churn, though. 22% of all jobs covered will change structurally, 39% of core skills will shift, and 41% of employers plan to cut headcount where AI can take over tasks. At the same time, 77% primarily plan upskilling rather than layoffs.
Which occupations are affected:
3. Reality Check: What Has Actually Happened So Far
Forecasts are one thing. The outplacement firm Challenger, Gray & Christmas counts how many US job cuts are actually attributed to AI, and in 2026, AI became the top stated reason for the first time:
In the first half of 2026, around 101,700 cuts were explicitly attributed to AI, roughly 23% of all announced reductions, and AI led the layoff reasons four months in a row. Cumulatively since 2023, it's about 173,600. The tech sector accounted for a good 139,000 cuts (+83% year-over-year), almost a third of all US reductions.
The best-known company cases with explicit AI reasoning:
The Klarna case deserves special attention. After the much-quoted 40% workforce reduction, CEO Sebastian Siemiatkowski admitted as early as May 2025 that they had gone too far ("We went too far"), and since 2026 Klarna has been hiring humans in customer service again because satisfaction suffered. It's one of the most honest statements in the entire debate.
4. The Junior Dip: AI Hits Entry-Level Workers
The clearest causal finding on AI's labor market effect so far comes from Erik Brynjolfsson and colleagues at Stanford, based on payroll data covering millions of US workers.
The result in one sentence:
Entry-level workers aged 22 to 25 lost 13% employment in heavily AI-exposed occupations since generative AI went mainstream, young software developers nearly 20% since late 2022, while experienced workers in the same occupations held steady or gained (Indeed puts the increase for older workers at 6 to 9%). The effect concentrates in automation-heavy occupations, not in those where AI augments people.
Other data sources paint the same picture. Per SignalFire, entry-level roles at big tech companies have collapsed by around 65% since 2019, and by 76% at startups; the CS class of 2025 was 45% less likely to land at a tech major than the class of 2022. Indeed measures a falling entry-level share of IT job postings while the senior share rises. Goldman Sachs most recently estimated AI's net effect at about minus 11,000 US jobs per month (as of June 2026, down from minus 16,000 in the spring), a regression estimate the bank itself flags as likely overstated; it attributes the decline mainly to temporary construction jobs around data centers, not to easing AI pressure.
5. Automation or Augmentation? What Usage Data Shows
The most interesting perspective comes from Anthropic's Economic Index, based on millions of anonymized Claude conversations. The central question is this:
Does AI replace the human (automation), or does it work alongside them (augmentation)?
For consumers, the two are nearly balanced. The enterprise view shows where things are heading, though:
Via the enterprise API, 75% of usage runs in automation mode. And cumulatively across all reports, 49% of occupations studied have had at least a quarter of their tasks done with Claude. The Cadences report from June 2026 adds the human side. Over a third of respondents expect AI to be able to handle most of their tasks within twelve months. But only 10% consider losing their own job likely, while over a third expect a younger colleague to lose theirs.
6. The Wage Premium: AI Skills Are the Strongest Salary Lever
While the loss side dominates headlines, the upside is growing quickly and quietly. For its AI Jobs Barometer, PwC analyzed over a billion job postings across 27 countries:
62% more pay for comparable roles requiring AI skills, and still rising. In some sectors the premium reaches 118%.
The PwC report also delivers the strongest counterargument to the pure job-loss narrative:
Heavily AI-exposed companies grew their headcount by 52% since 2018, weakly exposed ones by only 36%. And their productivity growth is 40% higher. So far, companies that use AI productively hire more people on balance, not fewer.
7. The New Jobs: 1.3 Million and a Clear Country Divide
According to LinkedIn data, AI has already created 1.3 million new jobs, with around 639,000 AI-related job postings in the US alone from 2023 to 2025. "AI engineer" is the fastest-growing role among young professionals for the second consecutive year. The share of AI skills in all job postings shows a clear country divide:
Singapore leads at 4.7%; Germany doesn't appear in the top 5. The story of the "prompt engineer" is a curious one. Career platforms report that the job title itself is disappearing from postings, while the skill migrates into roles like AI engineer as a requirement. The work remains; the job title consolidates.
8. Germany: 19% Have Already Cut Jobs
Several credible German surveys now exist. They measure different things and deliberately belong side by side, not merged:
The Bitkom number is the hardest one: 19% of companies have already cut jobs because of AI. The IAB figure (38% in occupations with high substitution potential) measures technical feasibility only, not actual replacement. Stepstone adds to the picture: in spring 2026, 36% of surveyed recruiters reported teams coming under review and a quarter reported layoffs, though Stepstone names general labor market pressure, not AI, as the main trigger. And searches for career changes out of office jobs rose over 100%, per Stepstone's "Arbeitsmarkt 2026" report. For how AI adoption in companies is evolving overall, see my AI agent statistics.
9. What Workers Think
According to Pew Research (5,273 US workers surveyed), 52% are worried about AI's future impact on their workplace, 32% expect fewer opportunities for themselves long-term, and only 36% feel hopeful about the development.
The most revealing number, though, sits in the Anthropic survey mentioned above:
Only 10% consider losing their own job likely, while over a third expect a younger colleague to lose theirs. This gap between self-assessment and assessment of others matches the Stanford finding from section 4 exactly. The disruption is perceived as real, just mostly for someone else.
10. Conclusion: Disruption Yes, Wipeout No
The 2026 data boils down to three sentences. First, actual AI job losses (173,600 attributed US cuts since 2023) remain far below the apocalyptic forecasts, and cases like Klarna show that overly aggressive automation gets corrected. Second, the effect is extremely unevenly distributed: entry-level workers in exposed occupations bear the brunt, while experienced and AI-skilled workers benefit, with a 62% wage premium. Third, the upside is growing measurably, from 1.3 million new AI jobs to above-average headcount growth at AI-exposed companies.
AI's labor market effect is less an extinction wave than a massive redistribution, between generations, skill sets, and companies. For the individual, the takeaway is as banal as it is well-documented. AI competence is currently the best-paid additional qualification on the market.
For how many people use AI in the first place, see my article on how many people use AI. And I dig deeper into the agent productivity debate in my AI agent statistics.
