Language as social technology, style, power and justice in algorithmic governance.

Abstract

This article examines the algorithmic stylization of language in responses mediated by Artificial Intelligence systems, focusing on the relationship between textual form, authorship, power, and communicative justice. It is based on the assumption that the way something is said is not a peripheral element of the utterance, but a decisive component in the production of meaning, the attribution of responsibility, and the practical orientation of the recipient. Grounded in a strictly theoretical study of a qualitative, bibliographic, and interpretive nature, the analysis brings together contributions from Philosophy of Language, Linguistics, Stylistics, Pragmatics, Discourse Analysis, and the ethics of technology. The article is organized around three central dimensions: authorship and responsibility, address and power relations, and aesthetics, readability, and standardization. The discussion shows that AI-generated responses may produce effects of unjustified authority when they display fluency, coherence, and an institutional tone without making their scope, limits, and recognizable grounds explicit. It also argues that address directly affects the recipient’s agency, either supporting understanding and action or reinforcing asymmetries and discursive dependence. At the formal level, the study contends that readability, organization, and expressive variation are conditions of access, whereas excessive standardization tends to generate tonal leveling, stylistic impoverishment, and the neutralization of contextual differences. It concludes that algorithmic stylization must be understood as a legitimate object of governance, since textual form is part of the duty to provide adequate information and participates in institutional responsibility in public and sensitive contexts.