Beyond the Screen: Exploring Privacy Boundaries through Automated User Profiling

Social Media Intelligence (SOCMINT) is widely used for gathering sensitive information about individuals, companies, or organizations, fueling potential attacks. This study highlights the inadequacy of current privacy protection measures and proposes an automated, sustainable approach to correlate user profiles, including homonyms and pseudonyms, solely through publicly available data. A sustainable method presents a solution to bypass API privatization and anti-scraping measures for data collection and formatting, emphasizing data correlation and the creation of cross-referenced dictionaries. Aggregated data enables indirect correlations, enhancing accuracy and relevance compared to existing methods. The purpose of this study is to create a self-learning knowledge base that addresses acronyms, homonyms, and pseudonyms while incorporating context to establish deeper connections between data. Comparisons will account for contextual relevance rather than relying solely on distances or metrics. Despite challenges posed by GDPR implementation and diminishing exploitable data, analysis of a database containing 7,600 accounts from major social platforms reveals over 90% accuracy in user-to-profile associations with 30% of homonyms and unpredictable pseudonyms or usernames. Notably, the study unveils the feasibility of retrieving information from private profiles solely through the application of dictionaries. The study demonstrates the limits of privacy on social networks, with users playing a passive role in the extensive dissemination of their data. Therefore, governance structures require innovative solutions to guarantee end-to-end protection and privacy.