Society & Work
How many jobs are lost to automation every day?
83 million jobs will be displaced by automation by 2027, the first net-negative wave in history
Roughly 20 jobs every minute.
Source: WEF Future of Jobs 2023; Oxford Martin School (Frey & Osborne 2013). View on dashboard →
Which jobs is automation displacing, and how fast?
Automation is eating jobs faster now. Robotics, AI, and RPA all contribute. WEF 2023: 83 million jobs displaced globally 2023-2027. BLS: 6.3 million US workers displaced from 2021-2023. Oxford Martin (2013) put 47% of US jobs at high automation risk within 20 years. That number is debated, but the direction is clear.
Job displacement over time
Job displacement from automation accelerated in the 2020s as large language models expanded automation from physical tasks into cognitive work, with WEF projecting 83 million jobs displaced globally between 2023 and 2027.
| Year | Rate | Jobs/day | Context |
|---|---|---|---|
| 2020 | 20/min | 29K | COVID drives automation acceleration |
| 2024 | 20/min | 29K | AI begins displacing knowledge workers; global rate modelled from US BLS data |
| 2027 (forecast) | 40/min | 58K | AI displaces knowledge work at scale |
| 2028 (forecast) | 48/min | 69K | LLM-powered cognitive automation expands |
Key automation displacement statistics
WEF 2023: 83 million jobs displaced globally 2023-2027; net loss of 14 million jobs
BLS: 6.3 million US workers displaced from jobs in 2021-2023
Oxford Martin (2013): 47% of US employment at "high risk" of automation within 20 years
McKinsey (2017) estimated 375-800 million workers globally could be displaced by automation by 2030
Jobs displaced vs. jobs created by automation, today
The net balance for 2023-2027 is -14 million globally. Both flows run simultaneously: each displaced role represents a human worker; each created role requires retraining and relocation.
The automation displacement wave: which work is vanishing and why
The displacement curve
Automation has displaced workers in manufacturing since the 1980s, but the current wave is qualitatively different: AI is now displacing cognitive work. Call centres, data entry, legal research, financial analysis, and content creation are all being automated by LLMs and specialised AI tools. The McKinsey Global Institute estimates 375-800 million workers face displacement by 2030. The wide range reflects uncertainty about AI adoption speed, regulatory responses, and the pace of new job creation in affected economies.
Why this wave is different
Previous automation waves primarily affected routine manual tasks, leaving non-routine cognitive work largely intact. The current wave of large language models and multimodal AI is the first to credibly threaten non-routine cognitive tasks: writing, analysis, coding, legal research, and creative work. This means the displacement risk has climbed the income and education ladder, reaching professional occupations that were previously considered "automation-proof." Workers with higher education are experiencing automation risk for the first time, and policy frameworks designed around manufacturing displacement are not well-suited to this new reality.
Research data
| Year | Finding | Value | Source |
|---|---|---|---|
| 2013 | Oxford Martin (Frey & Osborne): 47% of US employment at high risk of computerisation in next 10-20 years | 47 % US jobs at high automation risk | World Economic Forum |
| 2018 | WEF Future of Jobs 2018: 75M jobs displaced by 2022 | 75.0M jobs displaced projected 2018-2022 | World Economic Forum |
| 2020 | WEF 2020: 85M jobs displaced by 2025; COVID accelerates automation adoption | 85.0M jobs displaced projected 2020-2025 | World Economic Forum |
| 2023 | WEF 2023: 83M jobs displaced 2023-2027; first net negative projection (-14M) | 83.0M jobs displaced projected 2023-2027 | World Economic Forum |
| 2026 | WEF Four Futures for Jobs (Jan 2026): 4 AI/talent scenarios through 2030; 59% of the global workforce will need reskilling by 2030; 39% of core skills expected to become outdated | 59 % of global workforce needing reskilling by 2030 | World Economic Forum |
Key milestones
- 2013Oxford Martin School: 47% of US jobs at automation risk; landmark paper triggers global debate
- 2017McKinsey: 375-800M workers globally could be displaced by 2030
- 2020COVID-19 accelerates automation adoption: 43% of companies plan to reduce workforce due to technology (WEF)
- 2023WEF first negative net projection: 83M displaced vs 69M created 2023-2027
In perspective
Based on scaled BLS data, a worker somewhere loses their job to automation roughly every three seconds.
By this estimate, automation displaces roughly the equivalent of Belgium's entire workforce every year.
Every three seconds, a job somewhere changes permanently because of automation. In many cases, that work never comes back.
How the number is calculated
The live counter uses 20 jobs displaced per minute, derived from BLS displaced worker data (6.3M US workers displaced 2021–2023, approximately 3.15M/year) scaled to a conservative global estimate. For context, WEF 2023 projects 83 million displaced globally over 2023–2027 (≈ 40/min), which is the forecast ceiling; the counter uses the more conservative observed BLS-anchored baseline. The live counter accumulates displacement from midnight.
Sources: WEF - Future of Jobs Report 2025 - BLS - Displaced Workers Survey. Methodology →
Frequently asked questions
- How many jobs are being lost to automation per year?
- The WEF 2023 projects 83 million jobs displaced globally 2023-2027, or roughly 20.75 million per year. BLS data shows 6.3 million US displaced workers in 2021-2023. Automation's share of total job displacement is estimated at 15-30%.
- Which jobs are most at risk from automation?
- Oxford Martin (2013) identified clerk, data entry, accounts payable, telemarketing, and bank tellers as highest-risk roles. More recent analysis adds radiologists, paralegals, journalists, and customer service representatives due to AI language capabilities.
Why trust this data
The 83 million projection is from WEF's Future of Jobs 2023 report, based on employer surveys across 45 economies. US data comes from the BLS Displaced Workers Survey (published every two years). Oxford Martin School's foundational 2013 paper by Frey and Osborne has 10,000+ academic citations. These sources represent the gold standard for employment-automation research.
Explore related: Jobs gained from automation - Robots deployed - Automation statistics, and the live AnythingCounter dashboard.