Governance and academic integrity in the era of artificial intelligence: comparative documentary analysis in Latin American universities

Abstract

The IAGen study reveals transformations in the academic integrity guidelines of Latin American universities. This research compares thirteen (13) institutions from Peru, Colombia, Venezuela, Ecuador, and Costa Rica by analyzing their national policies, guides, guidelines, regulations, and frameworks for the use of AI, published between 2023 and 2026. A proprietary model was developed to analyze six dimensions: Governance and Leadership, Ethics, Equity and Transparency, Academic Integrity and Authenticity, Data Protection and Intellectual Property, Training and Competencies, and Continuous Improvement and Evaluation. This model is grounded in UNESCO's AI framework and the RAFTS model. Through a qualitative comparative document analysis, thirteen institutional documents published between 2023 and 2026 were coded across three levels of presence (absent, partial, complete), yielding an inter-coder reliability Cohen's Kappa of 0.78. The results identify three key findings: (i) a two-phase adoption process comprising pioneering first-generation institutions (high-level decrees) followed by a second operational wave characterized by specialized committees; (ii) a structural gap between ethical declarations and the operationalization of mandatory usage declaration mechanisms; and (iii) continuous updating as the least developed dimension. The study concludes that intraregional asymmetries demand collaborative frameworks and auditability mechanisms that extend beyond mere declarations, despite regional progress toward document governance.