Yichen Zhang, Lu Xian, and Florian Schuab, University of Michigan
Deepfake fraud—the use of AI-generated media to fabricate events for malicious purposes—threatens digital security, relationships, and public trust. This paper investigates individuals’ experiences with deepfake-driven scams through semi-structured interviews with seven participants. Most participants had limited prior exposure to harmful deepfakes and associated the technology with entertainment, underestimating its risks. During scams, they relied heavily on intuitive trust cues, such as familiar voices and social context, rather than verifying authenticity. Impersonations of moderately familiar individuals were more readily believed but less often verified, delaying deception detection. Emotional and relational harm, including confusion and damaged trust, often followed the incidents. Drawing on Protection Motivation Theory and Expectancy Violation Theory, this paper analyzes how perceived vulnerability, expectation alignment, and social familiarity shaped reactions, and offers design and policy recommendations for improving awareness, detection, and victim support.
Open Access Media
USENIX is committed to Open Access to the research presented at our events. Papers and proceedings are freely available to everyone once the event begins. Any video, audio, and/or slides that are posted after the event are also free and open to everyone. Support USENIX and our commitment to Open Access.
