Face off – a semiotic technology study of software for making deepfakes

Authors

  • Søren Vigild Poulsen Department of Language and Communication, University of Southern Denmark, Denmark

DOI:

https://doi.org/10.12697/SSS.2021.49.3-4.12

Keywords:

deepfakes, visual manipulation, semiotic software, face studies, digital culture

Abstract

Deepfakes, an algorithm that transposes the face of one person onto the face of another person in images and film, is a digital technology that may fundamentally alter our belief in visual modality and thus presents alarming consequences for an image-centric culture. Not only are these face-translations now so advanced that it is virtually impossible for people to tell that they are fake – this technology is also becoming accessible to laypersons who, with little or no computer skills, can use them for all kinds of purposes, including criminal intentions like revenge porn and identity theft. It is therefore timely and crucial to explore the semiotic potential of deepfakes.

This paper presents a semiotic technology perspective, i.e., the study of technology for meaning- making that is an emergent field in social semiotics, to report on findings from an ongoing study of how deepfake software is designed and used as a semiotic resource in erotic and political contexts. The paper advances the argument that the software is able to appropriate all signifiers of the face and their cultural history. Consequently, the semiotic operations of this technology prepare the ground for the problematic perspectives of synthetic facial imagery.

On this basis, the paper calls for a critical awareness of taking visual representations of current events at face value and considers how deepfake technology is embedded in unsound sharing practices of visual artefacts that tamper with the rich meaning potential of the face.

Downloads

Download data is not yet available.

Downloads

Published

2021-12-31

How to Cite

Poulsen, S. V. (2021). Face off – a semiotic technology study of software for making deepfakes. Sign Systems Studies, 49(3-4), 489–508. https://doi.org/10.12697/SSS.2021.49.3-4.12