Paper Number
ICIS2026-3235
Paper Type
Completed
Abstract
The commodification of deepfake technology led to a proliferation of interpretations of its uses. In online communities, where collective sense-making occurs, these interpretations (frames) may diffuse and converge across communities through shared membership. Understanding whether and how frames converge across communities, and toward which types of interpretations, is critical for assessing how collective sense-making shapes perceptions of and possibly future developments in deepfake technology. This study examines Reddit discussions following the release of a deepfake of Ukrainian President Volodymyr Zelensky in March 2022. We construct an inter-community network based on users’ posting activity and identify frames using a semi-supervised approach. We find that membership overlap fosters convergence, penalizing particularly harmful frames such as the generation of non-consensual sexual imagery. Furthermore, convergence toward specific frames largely depends on community specialization in discussing deepfake technology; while entertainment frames characterize these communities, they also spread both non-harmful frames, such as entertainment, and more problematic ones, such as disinformation.
Recommended Citation
Miotto, Marilu and Priante, Anna, "Making Sense of Deepfakes – How Online Communities Exchange Technological Frames on Reddit" (2026). ICIS 2026 Proceedings. 14.
https://aisel.aisnet.org/icis2026/social/social/14
Making Sense of Deepfakes – How Online Communities Exchange Technological Frames on Reddit
The commodification of deepfake technology led to a proliferation of interpretations of its uses. In online communities, where collective sense-making occurs, these interpretations (frames) may diffuse and converge across communities through shared membership. Understanding whether and how frames converge across communities, and toward which types of interpretations, is critical for assessing how collective sense-making shapes perceptions of and possibly future developments in deepfake technology. This study examines Reddit discussions following the release of a deepfake of Ukrainian President Volodymyr Zelensky in March 2022. We construct an inter-community network based on users’ posting activity and identify frames using a semi-supervised approach. We find that membership overlap fosters convergence, penalizing particularly harmful frames such as the generation of non-consensual sexual imagery. Furthermore, convergence toward specific frames largely depends on community specialization in discussing deepfake technology; while entertainment frames characterize these communities, they also spread both non-harmful frames, such as entertainment, and more problematic ones, such as disinformation.
Comments
24_Social_media