The 2013 automation predictions, and what happened next

Of the 135 jobs here that Frey and Osborne scored in 2013, they put 39 at a probability of computerisation of 0.90 or higher. By May 2025, 18 of those 39 employed more people than in May 2013, by BLS's count (BLS OEWS; change derived by Future History).

Data as of: Frey and Osborne (2013); BLS OEWS May 2013 and May 2025.

How this list was made

• Rows are the 135 jobs on our timelines that carry Frey and Osborne's published 2013 probability under the same occupation code, with BLS OEWS counts for May 2013 and May 2025 on one unbroken basis. Ordered by the 2013 probability, highest first.
• The answer counts jobs at 0.90 or higher; Frey and Osborne themselves called anything above 0.7 high risk.
• The change derived by Future History is computed from the two BLS counts.
• Every change here spans May 2021, when BLS moved to a model-based estimate that pools three years of survey data, so the counts at the two ends were made in different ways. A row is marked when its step from May 2020 to May 2021 is at least 3 times the size of its typical other one-year step and at least a quarter the size of the whole change shown. 37 of the 135 rows are marked.

What these numbers cannot tell you

• Frey and Osborne gave no date. They estimated what computers could technically do over "some unspecified number of years, perhaps a decade or two", not how many jobs would go, so this is not a test of a forecast they made.
• A job can grow for reasons that have nothing to do with automation, and shrink for others. The counts say what happened, not why.
• In May 2021 BLS changed how it estimates these counts (a model-based method using three years of survey data) and moved fully to the 2018 occupation codes, so part of a change across 2021 can come from the method rather than the work.
• BLS publishes these counts as yearly survey snapshots and cautions against reading them as a precise trend.

Left out because the occupation code changed

• Bus Drivers, Transit and Intercity: Frey and Osborne scored it under a 2010 occupation code that the 2018 codes renumbered, so the 2013 and latest counts may not cover the same workers.
• Computer Programmers: Frey and Osborne scored it under a 2010 occupation code that the 2018 codes renumbered, so the 2013 and latest counts may not cover the same workers.
• Educational, Guidance, and Career Counselors and Advisors: The 2018 occupation codes split, merged or renamed it, so the 2013 and latest counts may not cover the same workers.
• Fitness Trainers and Instructors: The 2018 occupation codes split, merged or renamed it, so the 2013 and latest counts may not cover the same workers.
• Light Truck Drivers: The 2018 occupation codes split, merged or renamed it, so the 2013 and latest counts may not cover the same workers.
• Special Education Teachers, Middle School: Frey and Osborne scored it under a 2010 occupation code that the 2018 codes renumbered, so the 2013 and latest counts may not cover the same workers.
• Stockers and Order Fillers: Frey and Osborne scored it under a 2010 occupation code that the 2018 codes renumbered, so the 2013 and latest counts may not cover the same workers.

Sources

• Frey and Osborne (2013), "The Future of Employment", appendix
• BLS OEWS national estimates, May 2013 and May 2025
• BLS, OEWS notice on the May 2021 estimates (new model-based method, MB3)
• BLS, 2010 SOC to 2018 SOC crosswalk (November 2017)
• Each row also names the exact sources for its own figures.

Explore these jobs by industry

• Admin & Office
• Manufacturing
• Construction
• Sales
• Transportation
• Engineering

← All trends and data stories