Henry Fabricio Carrillo Chacón, Fátima Lourdes Vega Barberán, Diego Alexander Cevallos Torres, María Lorena Torres Meza
The expansion of generative artificial intelligence challenges the capacity of traditional assessment to establish student authorship, understanding, and reasoning. This study analyzed the challenges these technologies pose to learning authenticity in Ecuadorian educational assessment and derived pedagogical guidelines consistent with national regulations. An integrative review with a systematized search and normative document analysis was conducted. The corpus preserved by the authors comprised 30 academic, institutional, and regulatory documents selected from 130 records using explicit criteria of relevance, recency, and full-text availability. The evidence was coded into six thematic categories. Results show that tasks focused exclusively on final products are particularly vulnerable to cognitive delegation; automated detectors do not provide sufficient evidence to attribute authorship; and the most consistent pedagogical response combines process traceability, situated performance, oral dialogue, feedback, and transparent disclosure of artificial-intelligence use. Ecuadorian regulations provide formative and contextual foundations for this redesign, although available literature does not establish how frequently these practices occur in the country's institutions. Learning authenticity therefore depends less on prohibiting all technological assistance than on aligning permitted assistance with learning outcomes and collecting complementary evidence of process and performance.