Year: 2025
Type: article
Abstract: This study draws on a dataset of 1394 flagged deepfakes from YouTube and TikTok annotated by the race, gender, profession and nationality of the ‘targeted’ individuals. The findings indicate that, by stark contrast to pornographic sites, mainstream deepfakes overwhelmingly feature individuals who are white, male, and—when in politics—authoritarian. A complementary qualitative analysis suggests that very few videos have a political edge, even when featuring politicians. Instead, it is proposed that the chief function of mainstream deepfakes is to be seen as a carnivalesque (apolitical) profanation of the faces of power. These findings challenge previous studies on the demographics of deepfake targets and broaden the scope of the current literature by shedding further light on how AI-assisted video doctoring is used beyond porn and disinformation.
Author: Carl Öhman
Institution: Uppsala University
Language: English
Subfield: Sociology and Political Science
Field: Social Sciences
Domain: Social Sciences
Open Access status: diamond
OpenAlex ID: W4411280844