ARTIFICIAL INTELLIGENCE IN DISTANCE EDUCATION: FORMATIVE ASSESSMENT, TEACHER MEDIATION, AND ETHICAL DILEMMAS IN CYBERCULTURE
DOI:
https://doi.org/10.55905/reiv6n2-011Keywords:
Artificial Intelligence. Distance Education. Formative Assessment. Pedagogical Mediation. Ethics of Singularity. Information Literacy.Abstract
The increasing integration of artificial intelligence (AI) tools in Higher Education and Distance Education (DE) has sparked intense debates regarding the future of pedagogical mediation, the integrity of student authorship, and the effectiveness of formative assessment practices. This article aims to analyze the pedagogical, epistemological, and ethical implications of incorporating AI into learning assessment and teaching practices within Virtual Learning Environments (VLEs). Methodologically, a qualitative, analytical-critical integrative literature review was conducted, guided by a structured protocol for document retrieval and screening. The analytical corpus comprises 18 benchmark scientific publications (six seminal foundational studies and 12 peer-reviewed articles indexed in the SciELO collection, primarily published between 2019 and 2026), which were submitted to thematic-categorical content analysis structured around four axes. The findings indicate that while educational data mining (EDM) and large language models (LLMs) demonstrate high efficacy in predicting student retention and generating immediate automated feedback, they suffer from "semantic blindness" and operate without human relational intentionality or empathy, necessitating active oversight under the "human-in-the-loop" model. Furthermore, the study reveals that the widespread reliance on synthetic text arises from the persistence of an atomistic ontology centered on controlling pre-programmed content ("learnification"), which demands the redesign of assessment tasks toward unique, empirical, and inquiry-based contexts. The paper critically addresses the rise of commodified "cyber-technicism" that seeks to replace faculty positions with standardized, off-the-shelf (prêt-à-porter) instructional packages, as well as the risks of the "algorithmic Pygmalion effect" and the existing gaps in information literacy and sector-specific regulation in Brazilian universities. It concludes that AI must function strictly as augmented intelligence serving critical emancipation, thereby repositioning formative assessment as an ethical engagement with otherness and student singularity.

