Abstract
What is the impact of automation on public sector employment? Using machine learning and natural language processing algorithms, this study estimates which occupations and agencies of the Brazilian Federal Government are most susceptible to automation. We contribute to the literature by introducing Bartik Occupational Tasks (BOT), an objective method used to estimate automation susceptibility that avoids subjective or ad hoc classifications. We show that approximately 20% of Brazilian public sector employees work in jobs with a high potential of automation in the coming decades. Government occupations with lower schooling and lower salary levels are most susceptible to future automation.