Analysis of automation and retraining opportunities for the Brazilian federal public service

Abstract

Context: The speed of technological advancement has raised concerns about the impact of automation on work. Previous studies have analyzed this impact on the private sector, but there is little research focused on the Brazilian public sector, which faces additional challenges in dealing with this issue. Objective: To estimate the impact of automation on intermediate-level federal public positions and indicate professional retraining paths for civil servants, in the face of technological changes. Methodology: A 4-step approach was used, involving (1) mapping of automation technologies, (2) extraction and review of the job duties of 142 intermediate-level public positions from a government document with the support of AI models, (3) assessment of the importance level and horizon of automation of each job duty by GPT-4 Turbo and (4) suggestion of professional retraining courses for each position. Results: A considerable potential for automation of the analysed positions was identified, with 72% of the job duties with different automation levels in an immediate or short-term impact horizon. However, also considering the frequency and importance of each job duty, a considerable part of the positions had an impact between 0.10-0.28 points on a scale of 0 to 1. The technologies with the most occurrences in the analysis of automation impact were Machine Learning, Smart Sensors and Process Digitization Systems. In the end, 236 professional retraining courses were suggested to better prepare employees for possible technological impacts.