Engenharia de software e inteligência artificial: uma análise de como o trabalho humano produz a IA

Abstract

This work aims to highlight the human being major role to create artificial intelligence systems. Through a selection of software engineering for design machine learning papers, I identify roles and workflow that demand human critical thinking in the process. As analytical tool, the monograph approaches the philosophy of technology of Álvaro Vieira Pinto, arguing the human being as active subject, whereas the machine as a mediation object of human labor and world. Because algorithms contain a purpose, AI cross-cutting techno-politics are explored, such as: AI is both embodied and material made from natural sources, labor and infrastructure; facade that occlude human labor behind AI; and imposition to infinitely accumulate more data. In that sense, I give emphasis in labeling crowdwork, that is responsible to give intelligence to AI, although it is a work in overexploitation conditions. With the references analyzed, it is discussed the importance of engineer awareness about production process and its policies. Finally, I mention suggestions, such as platform cooperativism, to think technique for the people.