Marginalized Bodies on Instagram

(In)Visibility, Control and Surveillance in Algorithmic Dispositive

Abstract

Artificial Intelligence (AI) emerged in the 20th century with the proposal of teaching machines to simulate human intelligence (McCarthy et al., 2006). In recent decades, this technology has evolved and come to play a central role in the management of both public and private life (Vicente, 2024). Algorithms, which organize and filter data, are an essential part of this process, controlling information flows and mapping user preferences (Gillespie, 2014). On social media, they operate through recommendation systems that personalize the content shown to users, creating the so-called "filter bubbles" (Pariser, 2011). In an algorithmic society, these systems can be understood as new discursive procedures, as they control, select, organize, and redistribute data, defining what appears, to whom, and when. Analyzing this dynamic through the lens of Foucauldian Discourse Studies, particularly by drawing on the concepts of discourse, power, control, and subject formation, this article problematizes the interference of AI in the circulation of discourse on Instagram, with a focus on content related to marginalized groups. The hypothesis is that social networks, beyond being spaces of expression, function as platforms that shape discursive visibility according to algorithmic and commercial logics. A clear example is the practice of shadowbanning, which operates as a contemporary mechanism of discursive exclusion. As a preliminary conclusion, it is observed that algorithms, by limiting the reach of certain discourses, reconfigure, under new technical conditions, the exclusion procedures described by Michel Foucault (1996).