Unmasking Cybercrime Strengthening Digital Identity Verification against Deepfakes 2026

Page 4 of 23 · WEF_Unmasking_Cybercrime_Strengthening_Digital_Identity_Verification_against_Deepfakes_2026.pdf

FIGURE 1 Anatomy of a deepfake attackBackground and motivation Face swapping is an AI-enabled form of impersonation and identity theft.01 The term deepfake refers to synthetic audio or visual media that convincingly imitates real people, created through machine learning methods (typically deep neural networks).2 This technology enables the creation of highly realistic but fabricated images, videos and audio recordings that are often hard to distinguish from authentic content.3 Originally developed for entertainment and creative applications, deepfake technology has increasingly been exploited for malicious purposes, including fraud, disinformation and impersonation.4One prominent technique within the domain of deepfakes is face swapping.5 Face swapping involves replacing the facial identity of an individual in an image or video with that of another person, while preserving expressions, movements and contextual realism.6 When misused, face swapping can serve as a sophisticated mechanism for identity theft, enabling attackers to engage in unauthorized activities under the guise of a victim’s likeness.7What are deepfakes and face swapping? Stolen or AI- generated identity documents are prepared.High-quality face-swapping tools are used to generate face-swapped media that matches submit- ted documents.Camera injection tools are used to feed synthetic video into live biometric checks.Cybercriminals bypass automatic identity verification. Unmasking Cybercrime 4
Ask AI what this page says about a topic: