Mapping Semantic Connections in Digital Content: The Role of the AI-Based SafeYou System in Supporting Multilingual Literacy and Lifelong Learning

In the digital age, access to information is being restructured not only through technological tools but also at cognitive and conceptual levels. The dominant logic of keyword-based retrieval in traditional search systems is increasingly giving way to models grounded in contextual and semantic processing. Considering the shortening of attention spans and the imperative of utilizing human labor more efficiently, this transformation is not merely inevitable but a functional necessity. In this context, the present study centers on a semantic matching approach powered by artificial intelligence and explores its potential to accelerate individual learning processes within the field of digital literacy. Through a conceptual analysis of a project output developed under this framework, the study aims to support users’ scientific, professional, and personal development via a system that prioritizes conceptual proximity over surface-level similarity in multilingual content. The research focuses on the “E-Center for Digital Literacy / Dijital Okuryazarlık E-Merkezi” platform, developed under the Erasmus+ project titled Digital Health Guidance Academy for Young Leaders. Within the platform, the SafeYou database contains 1,384 content records in multiple languages. These records were processed using the open-source multilingual-e5-large model, which transformed each title and summary into vector representations. Cosine similarity was then applied to compute semantic proximity between content items. As a result, for each entry, the 20 most semantically similar items were identified, yielding a total of 27,680 conceptual relationships. The system has demonstrated notable performance not only in terms of processing speed but also in conceptual richness. A multilingual matching process that would require an estimated 231 hours of human labor was completed by the AI model in approximately 29 seconds. This evaluation reveals that artificial intelligence can function not only as a tool for information retrieval but also as a conceptual guide and learning accelerator.