GENERATIVE ARTIFICIAL INTELLIGENCE LITERACY AND AUTHENTIC ASSESSMENT IN HIGHER EDUCATION: ACADEMIC INTEGRITY

Abstract

The emergence of generative artificial intelligence has reconfigured literacy paradigms and assessment vectors within the university ecosystem, challenging traditional notions of authorship, traceability of the formative process, and academic honesty. In this context, the present study examines the epistemic interdependencies among literacy in generative tools, the praxis of authentic assessment, and the dimensions of academic integrity in higher education. Methodologically, an empirical quantitative design with a multidimensional scope was adopted, encompassing exploratory, descriptive, and correlational components, based on a probabilistic sample of N = 603 institutional actors: n = 482 students and n = 121 faculty members affiliated with an Ecuadorian university. Data collection was conducted using a structured psychometric instrument composed of 30 items, whose analytical architecture operationalizes five core constructs: generative AI competence, academic use of generative environments, situated assessment, academic integrity, and learning traceability. The instrument was validated through expert judgment and showed adequate internal consistency across domains and high internal consistency at the global level. The results revealed mean scores above the midpoint of the scale in all domains. Academic integrity obtained the highest mean score, followed by generative AI literacy, whereas learning traceability showed the lowest value. The correlations were positive and statistically significant, particularly among authentic assessment, learning traceability, and academic integrity. It is concluded that these dimensions are interdependent and should guide the design of university assessments that are more situated, transparent, and verifiable.