Deepfake Detection and Multimedia Forensics: Investigating Synthetic Media, Image Forgery, and Video Manipulation in Cybercrime Cases

Abstract

The growing sophistication of synthetic media technologies, more so deepfakes, has created new challenges for cybercrime investigations and digital forensics. Previously, a tool for creative uses and entertainment, deepfakes are now extensively misused for harmful ends such as identity theft, fraud, disinformation campaigns, political manipulation, and harassment. This review collates existing literature on deepfake detection and multimedia forensics, with focus on image forgery, video manipulation, and multimodal analysis methods. Critical forensic technologies and uses from legacy metadata analysis and pixel-by-pixel analysis to AI-based solutions like Microsoft Video Authenticator and DARPA's MediFor initiative are investigated for their potential in verifying digital evidence. In spite of these advancements, investigators continue to grapple with ongoing challenges such as quick improvements in generative AI models, anti-forensic techniques, and issues of admissibility of AIgenerated evidence in courts. Future directions identify the unification of explainable AI, blockchain-supported provenance systems, IoT-supported real-time detection, and global collaboration as key strategies for enhancing forensic resilience. Through bridging technological innovation and legal and ethical considerations, multimedia forensics will continue to be the cornerstone to protecting digital trust and preventing the abuse of synthetic media in cybercrime.