ESRI Discussion Paper Series No.351 Do Digital Technologies Complement or Substitute for Human Labor?
Abstract
- We conduct difference-in-differences analyses on the effect of the introduction of AI on hours worked, employment, and the non-routineness of tasks based on an original online survey of workers in Japan.
- Our estimation results show that (a) the introduction of AI reduced hours worked, (b)increased the non-routineness of jobs in terms of the repetitiveness of tasks, and (c) required more regular employees, even though it had no significant effect on total employment.
- These results suggest that AI acts as both a complement to and a substitute for human labor.
Structure of the whole text(PDF-Format 1 File)
-
page31. INTRODUCTION
-
page72. DATA
-
page72.1. Data Collection
-
page92.2. Construction of Non-Routine Task Intensity Measure
-
page92.2.1. Survey Design for Quantification
-
page102.2.2. Calculation of Non-Routine Task Intensity
-
-
page112.3. Descriptive Statistics
-
page112.3.1. Histograms
-
page112.3.2. Crosstables
-
-
-
page123. THE MODEL
-
page164. ESTIMATION RESULTS
-
page164.1. Effect on Hours Worked
-
page174.2. Effect on Non-Routine Task Intensity
-
page174.3. Effect on Employment
-
page184.4. Estimation Results by Occupation
-
-
page205. CONCLUSION
-
page22REFERENCES
-
page24FIGURES AND TABLES
-
page40APPENDIX 1
-
page41APPENDIX 2
-
page42APPENDIX 3