ARTIFICIAL INTELLIGENCE AS AN EPISTEMIC AGENT: THE PREDICTIVE TURN AND THE TRANSFORMATION OF SCIENTIFIC RATIONALITY

Ramos-Zaga F. · Margens (2026) · DOI: 10.18542/rmi.v20i34.19581

Resumen

The expansion of artificial intelligence (AI) within scientific discovery has transformed the foundations of knowledge by replacing empirical verification with algorithmic prediction as the new criterion of truth. Accordingly, this article aims to critically analyze the ontological and epistemological status of AI, examining whether algorithmic prediction establishes a technoscientific determinism that redefines the notions of causality and scientific truth. The study reveals an epistemic mutation in which predictive efficacy replaces causal explanation, establishing a regime of operational predictability and a performative rationality where algorithmic models actively shape the very reality they seek to describe. The central finding indicates that contemporary science faces a paradox between precision and understanding, as the autonomy of computation erodes intelligibility. More broadly, this shift signals a horizon in which scientific rationality is redefined under a logic of anticipatory control, demanding a critical and responsible epistemology. © (2026), (Universidade Federal Do Para). All rights reserved. Esta obra está licenciada sob forma de uma licença Creative Commons Atribuição-Não Comercial 4.0 Internacional