INTELIGÊNCIA ARTIFICIAL NA EDUCAÇÃO: AUTORIA, INTEGRIDADE ACADÊMICA E AVALIAÇÃO EM XEQUE

Abstract

Generative Artificial Intelligence (GenAI) has become embedded in everyday academic practice before any consistent institutional or pedagogical response has been established. This theoretical essay, grounded in a narrative review of recent literature on Artificial Intelligence in Education (AIED), argues that GenAI does not merely represent a new pedagogical tool; rather, it simultaneously and structurally compromises three assumptions upon which formal education is organized. The first assumption is that written text constitutes evidence of one’s own thinking; the second is that learning requires the learner’s cognitive effort; and the third is that the evaluator is able to distinguish what the student knows from what the student has produced. These three assumptions are examined across conceptual, empirical, epistemological, and ethical dimensions, drawing on international and national normative frameworks as well as evidence concerning authorship, academic integrity, and cognitive atrophy. It is concluded that the uncritical incorporation of GenAI, without the field explicitly naming what is being compromised, is more dangerous than the tool itself.