Academic Discourse on AI in Europe and Latin America: A Comparative Analysis with IRaMuTeQ

Abstract

This study conducted a comparative analysis of academic discourse on AI in Latin America and Europe, based on 32 scientific articles published between 2020 and 2025 in ScienceDirect. Using a qualitative-comparative approach, IRaMuTeQ software   was used to generate a Labbe-type similarity tree, lexical concurrence networks, and descending hierarchical classification (DHC), and contrasting semantic configurations were identified. In Europe, a fragmented and diversified discourse predominates, with a tendency towards AI regulation, technical governance, and advanced development of machine learning and automation. In Latin America, on the other hand, a cohesive and pragmatic discourse is observed, with semantic nuclei associated with legal processes, public management, education, and social equity. These divergences, far from being merely linguistic, express asymmetry related to investment, technological maturity, and institutional policies; the Latin American region accounts for only 1.56% of global spending on AI (Jung, 2025), compared to a consolidated European ecosystem in research and development. It was determined that discourses on AI constitute situated socio-technical artifacts, whose form reflects both the opportunities and limitations of each region, and underscores the urgency of moving toward interregional cooperation mechanisms that enable inclusive, democratic, and contextualized technological advances.