The Impact of Artificial Intelligence on the Labor Market

I develop a new method to predict the impacts of a technology on occupations. I use the overlap between the text of job task descriptions and the text of patents to construct a measure of the exposure of tasks to automation. I first apply the method to historical cases such as software and industrial robots. I establish that occupations I measure as highly exposed to previous automation technologies saw declines in employment and wages over the relevant periods. I use the fitted parameters from the case studies to predict the impacts of artificial intelligence. I find that, in contrast to software and robots, AI is directed at high-skilled tasks. Under the assumption that the historical pattern of long-run substitution will continue, I estimate that AI will reduce 90:10 wage inequality, but will not affect the top 1%.

[1]  George A. Miller,et al.  WordNet: A Lexical Database for English , 1995, HLT.

[2]  Edward W. Felten,et al.  Linking Advances in Artificial Intelligence to Skills , Occupations , and Industries , 2017 .

[3]  David Autor,et al.  The Growth of Low Skill Service Jobs and the Polarization of the U.S. Labor Market , 2009, SSRN Electronic Journal.

[4]  Byron Anderson,et al.  Dictionary of Occupational Titles , 1993 .

[5]  JoAnne Yates,et al.  Structuring the Information Age: Life Insurance and Technology in the Twentieth Century , 2005 .

[6]  Andrew Janowczyk,et al.  Deep learning for digital pathology image analysis: A comprehensive tutorial with selected use cases , 2016, Journal of pathology informatics.

[7]  Daron Acemoglu,et al.  The Race between Man and Machine: Implications of Technology for Growth, Factor Shares, and Employment , 2018, American Economic Review.

[8]  Tom M. Mitchell,et al.  What can machine learning do? Workforce implications , 2017, Science.

[9]  Magne Mogstad,et al.  The Skill Complementarity of Broadband Internet , 2015, SSRN Electronic Journal.

[10]  Michael A. Osborne,et al.  The future of employment: How susceptible are jobs to computerisation? , 2017 .

[11]  Geraint Rees,et al.  Clinically applicable deep learning for diagnosis and referral in retinal disease , 2018, Nature Medicine.

[12]  Mark Johnson,et al.  An Improved Non-monotonic Transition System for Dependency Parsing , 2015, EMNLP.

[13]  Per-Anders Edin,et al.  Individual Consequences of Occupational Decline , 2019, SSRN Electronic Journal.

[14]  James H. Martin,et al.  Speech and language processing: an introduction to natural language processing, computational linguistics, and speech recognition, 2nd Edition , 2000, Prentice Hall series in artificial intelligence.

[15]  Richard B. Freeman,et al.  Who owns the robots rules the world , 2015 .

[16]  Guy Michaels,et al.  Robots at Work , 2015, Review of Economics and Statistics.

[17]  David Autor,et al.  Skills, Tasks and Technologies: Implications for Employment and Earnings , 2010 .

[18]  Shane Legg,et al.  Human-level control through deep reinforcement learning , 2015, Nature.

[19]  Frank Levy,et al.  Can Robots Be Lawyers? Computers, Lawyers, and the Practice of Law , 2016 .

[20]  Lynn Wu,et al.  The Employment Consequences of Robots: Firm-Level Evidence , 2019, SSRN Electronic Journal.

[21]  Edward N. Wolff,et al.  Trends in the Growth and Distribution of Skills in the U.S. Workplace, 1960–1985 , 1991 .

[22]  Richard S. Sutton,et al.  Temporal credit assignment in reinforcement learning , 1984 .

[23]  Tom M. Mitchell,et al.  What Can Machines Learn, and What Does It Mean for Occupations and the Economy? , 2018 .

[24]  A. Krueger.  How Computers Have Changed the Wage Structure: Evidence from Microdata, 1984-1989 , 1991 .

[25]  Daron Acemoglu,et al.  Robots and Jobs: Evidence from US Labor Markets , 2017, Journal of Political Economy.

[26]  Abhinav Vishnu,et al.  Deep learning for computational chemistry , 2017, J. Comput. Chem..

[27]  Razvan Pascanu,et al.  Sim-to-Real Robot Learning from Pixels with Progressive Nets , 2016, CoRL.

[28]  E. Brynjolfsson,et al.  The Productivity J-Curve: How Intangibles Complement General Purpose Technologies , 2018, American Economic Journal: Macroeconomics.

[29]  Daan Wierstra,et al.  Stochastic Backpropagation and Approximate Inference in Deep Generative Models , 2014, ICML.

[30]  Cliff T. Bekar,et al.  General purpose technologies , 2013 .

[31]  James Bessen.  BookIT IP Series - Learning by Doing: The Real Connection Between Innovation, Wages, and Wealth , 2015 .

[32]  D. Deming,et al.  The Growing Importance of Social Skills in the Labor Market , 2017 .

[33]  N. Fortin,et al.  Occupational Tasks and Changes in the Wage Structure , 2011, SSRN Electronic Journal.

[34]  M. Arntz,et al.  Revisiting the risk of automation , 2017 .

[35]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[36]  P. David.  The Dynamo and the Computer: An Historical Perspective on the Modern Productivity Paradox , 1990 .

[37]  A. Agrawal,et al.  Artificial Intelligence , Scientific Discovery , and Commercial Innovation ∗ , 2019 .

[38]  John Schulman,et al.  Concrete Problems in AI Safety , 2016, ArXiv.

[39]  Richard J. Murnane,et al.  Upstairs, Downstairs: Computers and Skills on Two Floors of a Large Bank , 2001 .

[40]  M. Kleiner,et al.  Occupational Licensing and Labor Market Fluidity , 2020, SSRN Electronic Journal.

[41]  J. Van Reenen,et al.  Has ICT Polarized Skill Demand? Evidence from Eleven Countries over Twenty-Five Years , 2014, Review of Economics and Statistics.

[42]  James Bessen,et al.  Automatic Reaction - What Happens to Workers at Firms that Automate? , 2019, SSRN Electronic Journal.

[43]  John Salvatier,et al.  When Will AI Exceed Human Performance? Evidence from AI Experts , 2017, ArXiv.

[44]  Francesca Cornaglia,et al.  Job Creation and Job Destruction , 2002 .