Inteligência artificial e mercado de trabalho: uma mudança perpétua?

Abstract

This study seeks to analyze how technology relates to the United States labor market. The main objectives are twofold: the first is to study how technological innovations, more specifically artificial intelligence, affect workers' wages in the United States. The second is to perform with SARIMAS models and neural networks to make a projection of wages for the next three years. In view of the changes in social and productive relations brought about by technology that have been observed in recent decades, together with the fundamental role of income in an economic system, this study is of great importance. To verify this relationship, a theoretical framework was used that categorizes jobs into types of non-routine or routine and cognitive or manual tasks. Based on this classification, a database was built and econometric and machine learning methods were adopted in the development of the study, in order to capture the trajectory of the categories of American wages It can be noted that the behavior for the absolute value of wages differs from the behavior observed for the value of wages as a proportion of GDP. These differences are clear and represent important aspects of the technological unemployment phenomenon. In general, the good predictive performance of the MLP neural networks stands out and the growing trajectory of each salary category for the short term, considering the effects of technology.