{"type":"script","doc_desc":{"producers":[{"name":"Reproducibility WBG","abbr":"DECDI","affiliation":"World Bank - Development Impact Department","role":"Verification and preparation of metadata"}],"prod_date":"2025-11-18","version":"1"},"project_desc":{"authoring_entity":[{"name":"Yan Liu","affiliation":"World Bank","email":"yanliu@worldbank.org"},{"name":"Shu Yu","affiliation":"World Bank","email":"syu2@worldbank.org"},{"name":"He Wang","affiliation":"World Bank","email":"hwang21@worldbank.org"}],"title_statement":{"title":"Reproducibility package for Labor Demand in the Age of Generative AI: Early Evidence from the U.S. Job Posting Data","idno":"RR_USA_2025_436"},"data_statement":"Some data is restricted and has not been included in the reproducibility package. For more details, please refer to the README file.","software":[{"name":"R","version":"4.4.0"},{"name":"Python","version":"3.12.3"},{"name":"Stata","version":"18.5 MP"}],"scripts":[{"title":"Reproducibility package for Labor Demand In The Shadow Of Generative Ai: Evidence From The U.s. Job Market","date":"2025-11","notes":"Computational reproducibility verified by Development Impact (DECDI) Analytics team, World Bank.","instructions":"See README in reproducibility package.","file_name":"RR_USA_2025_436","zip_package":"RR_USA_2025_436.zip","dependencies":"R: arrow, haven, dplyr, stringr, readxl, conflicted, ggplot2, lubridate, scales, tidyr, fixest, modelsummary, tibble, broom, colrspacem, gridExtra.\nPython: pandas, pyreadstat, pyspark.\nNo dependencies are used for Stata."}],"repository_uri":[{"name":"Reproducible Research Repository (World Bank)","uri":"https:\/\/reproducibility.worldbank.org"}],"production_date":"2025-11-18","abstract":"This paper examines the causal impact of generative artificial intelligence on U.S. labor demand using online job posting data. Exploiting ChatGPT\u2019s release in November 2022 as an exogenous shock, the paper applies difference-in-differences and event study designs to estimate the job displacement effects of generative artificial intelligence. The identification strategy compares labor demand for occupations with high versus low artificial intelligence substitution vulnerability following ChatGPT\u2019s launch, conditioning on similar generative artificial intelligence exposure levels to isolate\nsubstitution effects from complementary uses. The analysis uses 285 million job postings collected by Lightcast from the first quarter of 2018 to the second quarter of 2025Q2. The findings show that the number of postings for occupations with above-median artificial intelligence substitution scores fell by an average of 12 percent relative to those with below-median scores. The effect increased\nfrom 6 percent in the first year after the launch to 18 percent by the third year. Losses were particularly acute for entry-level positions that require neither advanced degrees (18 percent) nor extensive experience (20 percent), as well as those in administrative support (40 percent) and professional services (30 percent). Although generative artificial intelligence generates new occupations and enhances productivity, which may increase labor demand, early evidence suggests that some occupations may be less likely to be complemented by generative artificial intelligence than others.","geographic_units":[{"name":"United States of America","code":"USA"}],"keywords":[{"name":"Generative Artificial Intelligence"},{"name":"Technology Adoption"},{"name":"Labor Demand"},{"name":"Online Job Postings"}],"topics":[{"id":"O33","uri":"https:\/\/www.aeaweb.org\/econlit\/jelCodes.php?view=jel","vocabulary":"Journal of Economic Literature (JEL)","name":"Technological Change: Choices and Consequences \u2022 Diffusion Processes","parent_id":"O3"},{"id":" J23","uri":"https:\/\/www.aeaweb.org\/econlit\/jelCodes.php?view=jel","vocabulary":"Journal of Economic Literature (JEL)","name":"Labor Demand","parent_id":"J2"},{"id":" J21","uri":"https:\/\/www.aeaweb.org\/econlit\/jelCodes.php?view=jel","vocabulary":"Journal of Economic Literature (JEL)","name":"Labor Force and Employment, Size, and Structure","parent_id":"J2"}],"output":[{"type":"Working Paper","description":"Policy Research Working Papers (PRWP) 11263","title":"Labor Demand in the Age of Generative AI: Early Evidence from the U.S. Job Posting Data","uri":"http:\/\/documents.worldbank.org\/curated\/en\/099827011182513988","doi":"https:\/\/doi.org\/10.1596\/1813-9450-11263"}],"language":[{"name":"English","code":"EN"}],"disclaimer":"The materials in the reproducibility packages are distributed as they were prepared by the staff of the International Bank for Reconstruction and Development\/The World Bank. The findings, interpretations, and conclusions expressed in this event do not necessarily reflect the views of the World Bank, the Executive Directors of the World Bank, or the governments they represent. The World Bank does not guarantee the accuracy of the materials included in the reproducibility package.","license":[{"name":"Modified BSD3","uri":"https:\/\/opensource.org\/license\/bsd-3-clause\/"}],"contacts":[{"name":"Yan Liu","affiliation":"World Bank","email":"yanliu@worldbank.org"},{"name":"Reproducibility WBG","affiliation":"World Bank","email":"reproducibility@worldbank.org"}],"technology_requirements":"Runtime: 24 hours.","technology_environment":"Paper exhibits were reproduced on two systems with the following specifications:\n\u2022 Computer 1 - Stata:\n\u2013 OS: Windows 11 Enterprise, version 24H2\n\u2013 Processor: Intel(R) Core(TM) Ultra 7 165U (2.10 GHz)\n\u2013 Memory available: 32 GB\n\u2022 Databricks - R and Python:\n\u2013 Runtime: 16.4, Spark 3.5.2, Scala 2.12\n\u2013 Driver: Standard_D16ds_v4, 64GB, 16 Cores\n\u2013 Worker: Standard_D16ds_v4, 64GB, 16Cores","reproduction_instructions":"**1. Access the data:** Three datasets used by the reproducibility package are not included in it. Users need to gain access to all data to be able to run the code. See the README and Datasets description for details.\n**2. Run the Stata code:** Modify the file path in the do-file \"OccupationLevelData\" and run it.\n**3. Run the Python and R notebooks on Databricks:** The notebooks in the folder \"noteboom\/\" run on Databricks. Once they are placed in a Databricks workspace, they need to run in the following order: \"1 LightCast Data Aggregation.ipynb\", \"2 Prepare Data for Analysis.ipynb\", and \"3 Results.ipynb\".\n\nSince not all the data are included, the package includes the results produced in the reproducibility verification. These files can be used to review the results presented in the paper.","datasets":[{"name":"GenAI Exposure Scores by ISCO-08 classification  - Supplementary data for: Generative AI and Jobs: A global analysis of potential effects on job quantity and quality","note":"Source: Gmyrek, Berg, and Bescond (2023). Data was accessed on October 16, 2024. Data file: \"rawdata\/GBB-scores-2024-10-16.csv\".","access_type":"Data is publicly available and included in the reproducibility package.","uri":"https:\/\/pgmyrek.shinyapps.io\/AI_Data_Portal_Research\/","citation":"Gmyrek, P., Berg, J., Bescond, D. \n2023.\n\"GenAI Exposure Scores - Supplementary data for: Generative AI and Jobs: A global analysis of potential effects on job quantity and quality [Dataset]\".\nDownloaded from https:\/\/pgmyrek.shinyapps.io\/AI_Data_Portal_Research\/.\nAccessed October 16, 2024."},{"name":"Global Index of Occupational Exposure to Generative AI - Supplementary data for: Generative AI and Jobs - A Refined Global Index of Occupational Exposure","note":"Source: Gmyrek, P., Berg, J., Kami\u0144ski, K., Konopczy\u0144ski, F., \u0141adna, A., Nafradi, B., Ros\u0142aniec, K.,Troszy\u0144ski, M. (2025).\nData was accessed on June 15, 2025.\nData file: \"rawdata\/ILO2025_Final_Scores_ISCO08_Gmyrek_et_al_2025.xlsx\".","access_type":"Data is publicly available and included in the reproducibility package.","uri":"https:\/\/github.com\/pgmyrek\/2025_GenAI_scores_ISCO08\/blob\/main\/Final_Scores_ISCO08_Gmyrek_et_al_2025.xlsx","citation":"Gmyrek, P., Berg, J., Kami\u0144ski, K., Konopczy\u0144ski, F., \u0141adna, A., Nafradi, B., Ros\u0142aniec, K.,Troszy\u0144ski, M.\n2025.\n\"Global Index of Occupational Exposure to Generative AI - Supplementary data for: Generative AI and Jobs - A Refined Global Index of Occupational Exposure [Dataset]\".\nDownloaded from \"https:\/\/github.com\/pgmyrek\/2025_GenAI_scores_ISCO08\/blob\/main\/Final_Scores_ISCO08_Gmyrek_et_al_2025.xlsx\".\nAccessed June 15, 2025."},{"license_uri":"https:\/\/www.bls.gov\/bls\/linksite.htm","uri":"https:\/\/download.bls.gov\/pub\/time.series\/jt\/jt.data.2.JobOpenings (series: JTS000000000000000JOL)","name":"Job Openings and Labor Turnover Survey (JOLTS)","note":"Source: Bureau of Labor Statistics (BLS).\nData was accessed on July 5, 2025.\nData file: \"rawdata\/JOLTS.xlsx\".\n","access_type":"Data is publicly available and included in the reproducibility package.","citation":"Bureau of Labor Statistics.\nn.d.\n\"Job Openings and Labor Turnover Survey (JOLTS) [Dataset]\".\nAccessed from https:\/\/download.bls.gov\/pub\/time.series\/jt\/jt.data.2.JobOpenings.\nAccessed July 5, 2025."},{"name":"AI Complementarity - Supplementary data for: Labor Market Exposure to AI: Cross-country Differences and Distributional Implications","note":"Source: Pizzinelli, C., A. Panton, M. M. Tavares, M. Cazzaniga, and L. Li. Data was shared by the data owners following an email request to the authors of \"Labor Market Exposure to AI: Cross-country Differences and Distributional Implications\" (2023).\nData was accessed on November 15, 2024.\nData file: \"rawdata\/AIOE_CAIOE_theta_for_sharing.xlsx\".","access_type":"Data access was granted directly to the study authors by the data owners. It was obtained with a custom data license that does not allow for redistribution and it is not included in the reproducibility package.","citation":"Pizzinelli, C., Panton, A., Tavares, M. M., Cazzaniga, M., Li, L.\nn.d.\n\"AI Complementarity - Supplementary data for: Labor Market Exposure to AI: Cross-country Differences and Distributional Implications [Dataset].\"\nUnpublished data.\nAccessed November 15, 2024."},{"name":"Occupation description and skills - Supplementary data for Click, Code, Earn: The Returns to Digital Skills","note":"Source: Martins-Neto, A., Liu, Y., Khurana, S., Porraz Lopez, J. M. Data was shared by the data owners following an email request to the authors of \"Click, Code, Earn: The Returns to Digital Skills\" (2025).\nData was accessed on December 9, 2024.\nData file: \"rawdata\/onet_occ.xlsx\".","access_type":"Data access was granted directly to the study authors by the data owners. It was obtained with a custom data license that does not allow for redistribution and it is not included in the reproducibility package.","citation":"Martins-Neto, A., Liu, Y., Khurana, S., Porraz Lopez, J. M.\nn.d.\n\"Occupation description and skills - Supplementary data for Click, Code, Earn: The Returns to Digital Skills [Dataset].\"\nUnpublished data.\nAccessed December 9, 2024."},{"name":"Raw Job Postings","note":"Source: Lightcast.\nData accessed was purchased from the source and made available with a database connection enabled for the authors' Databricks workspace. Hence, there is no raw data file for this dataset. Data was extracted and aggregated from the database using the Databricks notebook \"1 LightCast Data Aggregation.ipynb\" and was saved into the file \"rawdata\/LC_USA_MonthlyPostings.csv\".\nData was accessed on July 5, 2025.","access_type":"Data access requires purchase and is not included in the reproducibility package.","citation":"Lightcast.\nn.d.\n\"Raw Job Postings [Dataset]\".\nDistributed via Lightcast Data.\nAccessed July 5, 2025."}]},"tags":[{"tag":"DOI"},{"tag":"Open Code"},{"tag":"Restricted Data"}],"schematype":"script"}