AN EXPERIMENTAL FORENSIC ANALYSIS OF SOCIAL MEDIA-BASED CYBER ATTACKS

Abstract

Abstract: Social media platforms have become a primary vector for cyber attacks, with threat actors exploiting user trust, platform features, and technical vulnerabilities to compromise accounts, steal credentials, and perpetrate fraud. This paper presents an experimental forensic analysis of social media-based cyber attacks, examining attack methodologies, digital evidence collection frameworks, detection techniques, and legal admissibility challenges. Drawing on recent case studies including the StealC v2 infostealer campaign targeting Facebook users (August 2025), Vietnamese threat actors leveraging fake job postings on LinkedIn (2025), and AI-powered social engineering tactics, the study proposes a comprehensive forensic investigation framework integrating machine learning-based detection with traditional digital forensic processes. Key findings indicate that behavioral analysis combined with natural language processing achieves superior detection accuracy for compromised accounts, achieving F1-scores of 0.89-0.93 in experimental evaluations. However, forensic investigators face mounting challenges from encryption, data volatility, cross-jurisdictional evidence collection, and the emergence of AI-generated synthetic content. The paper concludes with recommendations for enhancing forensic readiness on social media platforms and adapting legal frameworks to accommodate novel evidentiary sources. Keywords: Social media forensics, cyber attack investigation, digital evidence, compromised account detection, machine learning, phishing, social engineering, StealC malware