Estereótipos de gênero na interação com interfaces conversacionais dotadas de inteligência artificial: um estudo sobre o fenômeno e possibilidades de prevenção

Abstract

Conversational interfaces emerged to answer an increasing need for speed and mobility in accessing information and performing tasks. While this technology is increasing technically, particularly when powered by Artificial Intelligence, the fact that it is constantly designed with female names and voices by default raises questions about how this design decision reinforces negative gender stereotypes about women and provokes abusive interactions that, if naturalized, can hamper progress in gender equality. In order to understand this phenomenon and raise possibilities for intervention, this work aims to investigate the context around the development and application of conversational interfaces, unveiling how stereotypes are manifested in this technology, what may be behind this manifestation and what can be done to prevent it. The methodology used consisted of interacting, based on a predefined script, with conversational interfaces to identify the clues of gender incorporated in them, in addition to a bibliographic review that could bring explanatory hypotheses about the phenomenon and intervention possibilities.