The Roots and Routes of Deepfakes: Towards an Ontological, Typological and Sociotechnical Perspective

This paper offers a critical review of the evolving concept of deepfakes, tracing their roots in machine learning-based face-swapping technologies to their broader sociotechnical implications in contemporary society. While early research focused on technical detection and generation, the term ‘deepfake’ now broadly refers to identity deception, AI-driven or otherwise. We analyze how the barriers to deepfake creation have lowered due to accessible tools, and how public understanding has shifted from technical specificity to generalized concern. Drawing on literature from information systems (IS), media studies, and elsewhere, we develop a typology that categorizes deepfakes by intent, realism, technological accessibility, and sociotechnical impact. We also propose an initial ontology to clarify conceptual boundaries, distinguishing deepfakes from related terms such as “cheap fakes” and “synthetic media.” This interdisciplinary perspective contributes to IS discourse by framing deepfakes as dynamic artifacts embedded within complex sociotechnical systems.