Social media is an environment in which information is rapidly generated and disseminated through user interactions, allowing multiple threats—such as phishing attacks, the spread of misinformation, and distortion of information during disaster situ...
Social media is an environment in which information is rapidly generated and disseminated through user interactions, allowing multiple threats—such as phishing attacks, the spread of misinformation, and distortion of information during disaster situations—to emerge simultaneously. To address these threats, security and public safety domains have primarily relied on centralized automated systems, including rule-based filtering mechanisms, artificial intelligence (AI) models, and dedicated communication networks. However, such systems inherently depend on predefined rules, trained datasets, and fixed decision-making or dissemination structures, which limits their ability to respond effectively to rapidly evolving or non-structured threats. In particular, the social context reflected in how users perceive and react to content is often insufficiently incorporated into automated detection and response processes, constituting a critical limitation of automation-centered approaches.
To address this limitation, this study proposes the structural feasibility of a detection and response framework that complements centralized automated systems by incorporating collective intelligence–based user participation data derived from social media environments. Crowdsourced data—such as user reports, ratings, comments, and reactions—represent experience-based information accumulated through actual usage and have the potential to capture early risk signals and shifts in social perception that automated systems may fail to detect in a timely manner.
This study adopts the principle that user participation data should be utilized not as a substitute for automated decision-making, but as supplementary evidence supporting detection and response judgments. The proposed framework emphasizes reliability by incorporating mechanisms such as user credibility assessment and convergence analysis of multiple user judgments. In addition, to address potential false positives and malicious interference inherent in participatory systems, the framework includes procedures for filtering repetitive reports and anomalous user behaviors. Beyond the detection phase, the framework is designed as a cyclical and staged response process encompassing real-time alerts, temporary access restrictions, information disclosure, and the integration of user feedback.
Rather than implementing an operational system or conducting quantitative performance evaluations, this research adopts a scenario-based analytical approach to examine the applicability and structural validity of the proposed framework across security threat scenarios and public safety and disaster response contexts. Through this analysis, the study demonstrates that centralized automated systems and user participatory mechanisms are not mutually exclusive, but can coexist as complementary components, preserving their respective strengths while collectively enhancing situational awareness and decision-making flexibility.