INTELIGÊNCIAS ARTIFICIAIS GENERATIVAS E VIÉS ALGORÍTMICO: UM NOVO PARADIGMA SEMIÓTICO

GENERATIVE ARTIFICIAL INTELLIGENCES AND ALGORITHMIC BIAS: A NEW SEMIOTIC PARADIGM

Abstract

The article discusses, from the perspective of discursive semiotics, the mechanisms of meaning production and the phenomenon of bias in images generated by generative artificial intelligence (AI) systems. Drawing on the Greimasian model of meaning generation, it analyzes how algorithms and databases, when processing large volumes of information, reproduce cultural stereotypes and hierarchize forms of life. The study adopts an experimental approach using controlled prompts to observe how such systems figurativize regional and gender identities, thereby consolidating hegemonic representations. By recognizing that algorithmic bias is constitutive of the generative logic, the article offers insights for a critical semiotic reflection on the ways in which AI participates in the symbolic construction of the world, calling for new forms of reading and ethical responsibility toward technologies of meaning.