Lenguaje como sistema predictivo: hacia una perspectiva semiótico-informática del discurso

Abstract

This paper proposes a theoretical framework for exploring contemporary conditions of knowledge production and discursive practices in environments mediated by artificial intelligence (AI), as well as some broader implications of its use. It argues that algorithmic mediation introduces a predictive logic of language that reconfigures forms of discursive production. On this basis, the paper outlines a semiotic-computational perspective for the study of AI-mediated discourse. The central problem lies in the limited articulation between conceptions of language as a predictive system in computational fields and as a social practice in discourse studies, which hinders critical examination of these technologies, particularly in academic contexts. Adopting a theoretical and transdisciplinary approach, the study draws on sociocultural perspectives on literacies and discourse, together with contributions from computational linguistics and computer science. Through conceptual analysis, it examines the technical, structural, and functional principles of AI-based text generation systems and relates them to a semiotic understanding of meaning-making in digital environments. It also contrasts language as a sociotechnical system with language as situated practice. The paper ultimately proposes a conceptual framework that integrates the sociotechnical dimension of generative language models with discourse and literacy studies while highlighting implications for academic writing, authorship, and digital literacy practices.