Episteme of Techné and Generative Artificial Intelligence

Foundations, Functioning, and Ethical Implications

Abstract

This article presents a critical analysis of Generative Artificial Intelligences (GAIs), emphasizing their statistical foundations, computational architectures, and modes of textual and multimodal operation. It argues that understanding such technologies requires not only technical familiarity but also reflection on the historical, epistemological, and political conditions that guided their development. To this end, the concept of episteme of techné is proposed, aiming to make explicit the theoretical, metatheoretical, and socio-historical foundations that sustain the emergence of GAIs, thus allowing the identification of historical markers and political choices that determined their current configuration. This proposal dialogues with the Foucauldian notion of materialism, understood as an analysis of the materiality of bodies, discourses, and dispositifs. Beyond the conceptual dimension, ethical, environmental, and political implications are discussed, such as the massive energy consumption of datacenters, the urban gentrification associated with the installation of such infrastructures, and the concentration of power in a few global corporations. The goal is to provide a reference framework that enables researchers, educators, and professionals to critically understand GAIs, aligning their adoption with principles of human agency, diversity, sustainability, and transparency, as recommended by UNESCO’s Guide for Generative AI in Education and Research.