ARTIFICIAL INTELLIGENCE AS A TOOL FOR AUTOMATIC LEARNING ASSESSMENT

Abstract

The incorporation of computational tools in educational contexts has redefined evaluative processes for more than a decade, generating tensions between automation and professional teaching judgment. This article comparatively analyzes artificial intelligence (AI) and traditional formative assessment as instruments for measuring learning, investigating their viabilities, limitations, and implications in Brazilian schools. Lemos and Maissiat recognize that the use of educational technologies constitutes a latent necessity, adding values of technological advances and encouraging student teaching and learning; however, Brazilian reality indices reveal substantial disparities in network access, exposing mismatches of needs in opposite directions in public schools. Through bibliographic research grounded in Gil (2022) and Frainer (2022), the study examines techno-pedagogical possibilities and ethical barriers of evaluative automation. The general objective consists of critically analyzing the potential and limitations of artificial intelligence in automatic assessment, comparing it with traditional formative practices in real school contexts. It concludes that AI, without complementation by reflective formative practices mediated by educators, does not guarantee evaluative quality or educational equity, requiring permanent teaching mediation and functioning of technologies as circumscribed auxiliaries. Therefore, responsible implementations demand prior clarity about educational purposes, critical vigilance over effects on vulnerable populations, and priority investment in critical teacher training on technologies.