Information Sustainability Beyond Digital Access: Evidence of Machine Learning in Local Media Ecosystems in Ecuador

Information sustainability has emerged as a key dimension of social sustainability, as it supports equitable access to information, informed opinion formation, and institutional engagement. However, the expansion of digital platforms does not necessarily guarantee inclusive information ecosystems, particularly in local contexts characterized by structural inequalities. This study examines information sustainability and the digital divide by identifying media consumption profiles within a territorial context in Ecuador. Using data from a survey conducted in the province of Imbabura with 1,784 observations, a hybrid methodological approach combining cluster analysis and Random Forest (RF) algorithms was applied. Audience profiles were identified and validated based on media consumption patterns, levels of digitalization, and institutional engagement. The results reveal four distinct audience profiles with different levels of digital integration and institutional linkage. Findings indicate that the intensity and diversity of media consumption play a more decisive role than mere technological access. Digital access alone is insufficient to ensure information sustainability or foster institutional opinion formation; instead, differences in exposure, usage intensity, and media habits shape audience engagement. These findings highlight the need for segmented, territory-based communication strategies to strengthen information sustainability, reduce the digital divide, and reinforce the role of university media within local media ecosystems.