About the Data

Technological Substitution of Skills

Since Autor, Levy and Murnane described “the skill content of technological change” in 2003, interest has grown in understanding jobs as bundles of skills (or “tasks”), and how technology can impact demand for these various bundles. Frey and Osbourne gained fame and notoriety when they produced “computerization” (or “automation”) probabilities for occupations in 2013 and 2017. The salience of a task- or skills-based approach is clear – and occupational scores help to advance our understanding of how technology-substitution task-bias might affect our industries, industries, and sub-populations by race and gender.

However, a transparent method and tool for predicting the risk posed by automation exposure is not readily available. Current tools are behind paywalls and rely on proprietary or difficult-to-replicate or-update methods.

The Role of Occupational Automation Exposure Scores

Simply, occupational scores of automation exposure answer “what jobs have tasks that are the most exposed to automation technologies?”. The LMI Institute, in partnership with the Council for Community and Economic Research (C2ER) produces Automation Exposure Scores derived from intensity scores of selected O*Net Abilities, Work Activities, and Work Context characteristics associated with technological substitution. These scores help education, labor market information, and economic and workforce development officials identify the occupations and tasks that are the most likely to be most complemented, or substituted, by automation technologies. As Industry 4.0 technologies like artificial intelligence and robotics become more widely adopted by employers, it will become increasingly important that the workforce of the future can operate and cooperate with the technologies of the 21st century.

The Automation Exposure Scores allows analysts to identify industries and occupations that are changing or likely to change due to automation and opportunities to assess potential for skill and career transitions for threatened occupations. Industry, educators, and the civic sector can collaborate on delivering solutions that provide at-risk industries and occupations with the business services, training, and career transition-services necessary to support their resiliency and growth. While automation threatens to replace workers, it also presents opportunities for regions to embrace new and emerging occupational skillsets that provide workers with family-supporting wages and increase the resiliency of key regional industries.


The LMI Institute’s Automation Exposure Scores measure the intensity score of selected O*Net characteristics associated with technological substitution. This data allows analysts to identify occupations with a high task-exposure to automation technologies using methods that are updateable and replicable. To produce these estimates, the LMI Institute synthesizes intensity scores from sixteen O*Net Abilities, Work Activities, and Work Context characteristics using methods developed by Dr. Steve Hine[1] to adapt analyses conducted by Acemoglu and Autor 2011.[2] Both these methods were adapted from analyses conducted by Autor, Levy, and Murnane 2003[3] using Dictionary of Occupational Title characteristics.

Occupation-level automation exposure scores are created by distilling scores from 923 detailed O*Net occupations into scores for 737 SOC occupations that enable comparison with BLS Occupational Employment and Wage Statistics (OEWS) and Current Population Survey data.

The original study utilized selected sixteen O*Net work characteristics most associated with job-loss to automation to develop occupational automation exposure scores. This method produces Automation Exposure Scores that are updateable and transparent. Scores derived from the 18,300 tasks associated with the 800+ O*Net occupations would be too burdensome to accurately review and characterize on an ongoing basis especially as tasks are constantly evolving as technological change occurs. Instead, the sixteen work characteristics utilized in these scores are significantly statistically associated with changes in occupational employment and skill-composition resulting from adoption of automation technologies.

Cognitive Analytical and Cognitive Interpersonal scores indicate reduced exposure to automation. Routine Cognitive and Manual, and Nonroutine Manual scores indicate increased exposure to, and potential substitution, by automation technologies.

[1] Link Hine paper from MN Econ Quarterly

[2] Link paper Acemoglu and Autor 2011

[3] Link paper Autor, Murnane, and Levy 2003

Comparability with Alternative Automation Exposure Scores

The Automation Exposure Scores produced by the LMI Institute provide a replicable, updateable, and transparent method for determining occupation and task-level automation exposure. While other methods exist to determine automation exposure, these methods are often not designed to be applicable to regions and are produced with non-transparent methods that are impossible to replicate and update as the impact of automation technologies on the demand for skills becomes more certain. The methods utilized by the LMI Institute have communities in mind – including ensuring that public officials understand how the estimates are produced and updated.

The LMI Institute has conducted research to compare and validate its results with alternative scores of occupation-level automation exposure. LMI Institute staff compared automation exposure scores for the occupation “Roof Bolters, Mining” across four-competing measures. The LMI Institute’s automation exposure score-method suggests that this occupation has a high exposure to automation technologies.

Automation Exposure Score for Roof Bolters, Mining
LMI Institute 1
Felten, Raj, Seamans[1] 0.61
Frey and Osborne[2] 0.49
Webb[3] 0.37
McKinsey Future of Work[4] 0

When reviewing this occupation and related skills, using data from the BLS Occupational Employment and Wage Statistics and technical literature specific to the coal mining industry, LMI Institute staff validated the high exposure of this occupation to automation. Roof Bolters, Mining is a hazardous occupation that requires no education and little on-the-job training. Roof bolting machines that use predictive maintenance, automated environment-responsive and safer drilling techniques, and have duel-purpose applications are becoming increasingly common, decreasing the demand for operators.

[1] Link paper

[2] Link paper

[3] Average of score for Robotics and AI automation exposure. Link data source / paper

[4] As published by the Brookings Institution. Link data source / paper

Preferred Citation

LMI Institute. Automation Exposure Score. 2021. link

Data Requests and Other Questions

Please use our website or direct requests for data to the LMI Institute at 703-522-4983 or link.

Copyright Policy

Each issue of the Automation Exposure Score and related products is copyrighted. Printing, transferring into computer-readable format, or otherwise reproducing an entire Index report or any part thereof for sale is expressly prohibited unless written permission is obtained from the LMI Institute. News media, however, are permitted to use LMI Institute Automation Exposure Scores in editorial form in both paper copy and on the Internet, and are permitted to reproduce tables in part to illustrate text, provided appropriate credit is given to the LMI Institute. LMI Institute member states are permitted to reproduce data for their respective state and the nation and develop unique products using the data. Any organization republishing or repurposing the data in any form must provide appropriate credit to the LMI Institute by using the preferred citation listed on this page.

Users are granted no other reproduction rights without a negotiated Data-User License Agreement (DULA).

Any questions about copyright policy or reproduction rights should be addressed to the LMI Institute.