Chất lượng bằng chứng 4.75/5
Điểm đánh giá tám chiều theo tiêu chí chất lượng . Mỗi chiều được chấm từ 1–5.
- D1 Cơ sở nguồn
- 4/5
- D2 Uy tín nguồn
- 5/5
- D3 Tính toán
- 5/5
- D4 Sự không chắc chắn
- 5/5
- D5 Phạm vi
- 5/5
- D6 Văn bản
- 4/5
- D7 Sự trung thực về nhận thức
- 5/5
- D8 Tính đầy đủ của lưu ý
- 5/5
Không một con số đơn lẻ nào trả lời được ‘xác suất robot lấy đi việc làm của bạn là bao nhiêu’, bởi nghiên cứu tiệm cận nhất chưa bao giờ được xây dựng để tạo ra con số ấy. Bằng chứng mạnh nhất là nghiên cứu của Acemoglu và Restrepo về các vùng đi làm ở US từ 1990 đến 2007, cho thấy mỗi robot công nghiệp tăng thêm trên một nghìn lao động đã làm giảm tỷ lệ việc làm-trên-dân số của một khu vực địa phương khoảng hai phần mười điểm phần trăm và tiền lương khoảng bốn phần mười phần trăm. Đó là các độ co giãn trên toàn khu vực. Chúng mô tả cách toàn bộ thị trường lao động của một vùng dịch chuyển khi robot xuất hiện, bao gồm cả những người lao động bị đẩy sang các ngành khác và tác động lan tỏa lên những công việc chưa từng được tự động hóa. Biến điều đó thành một xác suất trọn đời cá nhân sẽ đòi hỏi những giả định mà nghiên cứu không đưa ra, nên câu trả lời trung thực là một con số theo từng người lao động có thể bảo vệ được là không tồn tại.
Mặt triển khai là có thật và đang tăng. International Federation of Robotics cho biết ngành sản xuất US đạt 307 robot trên 10,000 lao động vào năm 2024, tăng từ 285 của hai năm trước và là một phần của mật độ toàn cầu đã tăng gấp khoảng đôi trong bảy năm. Nhưng mật độ là thước đo có bao nhiêu cỗ máy được lắp đặt, chứ không phải bao nhiêu người bị thay thế. Robot cũng làm tăng sản lượng, hạ giá, và tạo ra công việc bảo trì cùng kỹ thuật, đó là lý do vì sao chính những thập niên tự động hóa gia tăng ấy lại trùng với dự báo của Bureau of Labor Statistics về 6.7 triệu việc làm US nhiều hơn giữa 2023 và 2033. Một cỗ máy có thể làm một phần công việc không giống với một cỗ máy xóa sổ chính người đang giữ công việc đó, và khoảng cách giữa hai điều này là nơi gần như toàn bộ sự bất định trú ngụ.
Rủi ro tập trung ở đâu thì rõ ràng hơn là nó lớn đến mức nào. Các hiệu ứng đo được cụm lại ở các vùng sản xuất ô tô và ở công việc sản xuất theo lề thói cùng công việc di chuyển vật liệu; nghề thủ công lành nghề, công việc chăm sóc và hầu hết các vai trò chuyên môn cho thấy rất ít phơi nhiễm robot trực tiếp, và một số còn đối mặt với tình trạng thiếu lao động thay vì dư thừa máy móc. Làn sóng kho bãi, hậu cần và robot hình người mới hơn của thập niên 2020 đang tăng nhanh nhưng quá mới để có được ước tính dịch chuyển đáng tin cậy, và những tranh chấp nổi bật như cuộc đình công về tự động hóa cảng 2024-2025 cho thấy sự xáo trộn là có thật ngay cả khi không có phép tính. Với một xu hướng triển khai mạnh mẽ, một hiệu ứng đo được lên việc làm có thật nhưng khiêm tốn, và không có con đường nào từ cả hai điều đó dẫn tới một xác suất theo từng người, mục này mang cờ no_reliable_estimate. Nỗi sợ là chính đáng; một con số trọn đời đơn lẻ thì không.
Sổ cái xác nhận
Mỗi con số dưới đây là những gì mỗi nguồn đã báo cáo, cùng với trích dẫn nguyên văn mà chúng tôi dựa vào và cách chúng tôi đưa ra con số của mình. Nhấp vào bất kỳ liên kết nào để xác minh trực tiếp.
-
[1] National Bureau of Economic Research / Journal of Political Economy 128(6) — Robots and Jobs: Evidence from US Labor Markets
Robots and Jobs: Evidence from US Labor Markets- Thống kê
One more robot per thousand workers reduces the employment-to-population ratio by about 0.18-0.34 percentage points and wages by 0.25-0.5 percent (preferred point estimates: ~0.2 pp and ~0.42%); the arrival of one robot in a local labor market is associated with roughly 3.3-6 fewer workers employed there- Trích đoạn
“"We estimate large and robust negative effects of robots on employment and wages across commuting zones. [...] one more robot per thousand workers reduces the employment to population ratio by about 0.18-0.34 percentage points and wages by 0.25-0.5 percent." ”
- Dữ liệu nguồn từ
- 2020-06-01
- Truy cập
- 2026-07-29
- Tính toán
- Acemoglu & Restrepo, "Robots and Jobs: Evidence from US Labor Markets," Journal of Political Economy 128(6), 2020 (NBER Working Paper 23285). The design is a commuting-zone quasi-experiment exploiting variation in industry exposure to industrial robots over 1990-2007; the published version's preferred point estimates are a 0.2-percentage-point fall in the employment-to-population ratio and a 0.42% wage decline per robot per thousand workers. The authors' own summaries put the local effect at roughly six fewer workers per robot in the immediate commuting zone and about 3.3 fewer nationally once trade and product-market spillovers are netted out. Crucially, these are local-equilibrium employment elasticities, NOT a per-worker lifetime displacement probability: they measure how a region's employment moves with robot adoption, including workers displaced into other sectors and knock-on effects on non-robot jobs. There is no defensible arithmetic path from this elasticity to "the odds a given US adult is replaced by a robot," which is why this entry carries no_reliable_estimate. The estimate is also period-specific (industrial robots, pre-2008) and does not cover warehouse, logistics, or service robots deployed since.
- Độc lập
- Seminal quasi-experimental study using International Federation of Robotics installation data mapped to US commuting zones. Independent from the task-exposure and survey sources.
-
[2] International Federation of Robotics (World Robotics 2025) — Robot Density Surges in Europe, Asia and the Americas
Robot Density Surges in Europe, Asia and the Americas- Thống kê
US manufacturing robot density reached 307 units per 10,000 employees in 2024, ranking 8th worldwide (up from 285 in 2022)- Trích đoạn
“"The United States ranks 8th worldwide with 307 units per 10,000 employees." ”
- Dữ liệu nguồn từ
- 2026-04-08
- Truy cập
- 2026-07-29
- Tính toán
- IFR World Robotics 2025 (data year 2024). Robot density — the operational stock of industrial robots per 10,000 manufacturing employees — is the standard measure of how far automation has actually been deployed. US density of 307 (2024), up from 285 in 2022, is used here only to establish that robot adoption is large and rising, which is the real trend underneath the fear. Density is a deployment measure, not a displacement rate: a higher robot count does not translate into a fixed number of jobs eliminated, because robots also raise output, lower prices, and create maintenance/engineering roles. No per-worker probability is derivable from a density figure.
- Độc lập
- Industry census of installed robots; the same IFR installation data underlies the Acemoglu & Restrepo identification strategy but is used here purely for the deployment trend.
-
[3] Bureau of Labor Statistics — Employment Projections: 2023-2033
Employment Projections: 2023-2033Xem tất cả 2 mục Likelier trích dẫn nguồn này →
- Thống kê
Total US employment projected to add 6.7 million jobs (0.4% annual growth) from 2023 to 2033, reaching 174.6 million- Trích đoạn
“"The U.S. economy is projected to add 6.7 million jobs from 2023 to 2033 [...] Total employment is projected to increase to 174.6 million and grow 0.4 percent annually." ”
- Dữ liệu nguồn từ
- 2024-08-29
- Truy cập
- 2026-07-29 · bản sao lưu trữ
- Tính toán
- BLS Employment Projections, 2023-2033 release (USDL-24-1776). Net job growth of 6.7 million through 2033 is a direct counterweight to headline claims of mass robotic job elimination: BLS incorporates historical technology-adoption and productivity trends and still projects net growth, not net loss. NOTE: bls.gov/news.release/ecopro.nr0.htm is a rolling endpoint that now serves the newer 2024-2034 projections, so archive_url pins the dated 2025-05-30 Wayback snapshot of the 2023-2033 release from which this figure and excerpt are drawn. This does not preclude heavy churn or concentrated losses in specific occupations and regions, but it bounds the population-level story. No per-worker displacement probability is published.
- Độc lập
- Official US government statistical agency using proprietary macroeconomic modelling. Fully independent from the academic and industry sources above.