Characterizing the MrDeepFakes Sexual Deepfake Marketplace

Authors: Catherine Han, Anne Li, Deepak Kumar, Zakir Durumeric

Published: 2024-10-14 21:25:42+00:00

Comment: To appear in USENIX Security Symposium 2025

AI Summary

This paper systematically characterizes MrDeepFakes, a prominent online marketplace for sexual deepfakes, by analyzing its economics, target demographics, and deepfake creation discussions. The research uncovers lax enforcement of rules, previously undocumented attacker motivations, and novel tactics for acquiring resources to create sexual deepfakes. It highlights the platform's rapid growth and the community-driven solutions to deepfake creation challenges.

Abstract

The prevalence of sexual deepfake material has exploded over the past several years. Attackers create and utilize deepfakes for many reasons: to seek sexual gratification, to harass and humiliate targets, or to exert power over an intimate partner. In part enabling this growth, several markets have emerged to support the buying and selling of sexual deepfake material. In this paper, we systematically characterize the most prominent and mainstream marketplace, MrDeepFakes. We analyze the marketplace economics, the targets of created media, and user discussions of how to create deepfakes, which we use to understand the current state-of-the-art in deepfake creation. Our work uncovers little enforcement of posted rules (e.g., limiting targeting to well-established celebrities), previously undocumented attacker motivations, and unexplored attacker tactics for acquiring resources to create sexual deepfakes.


Key findings
The MrDeepFakes platform experienced significant growth post-2021, hosting 43K sexual deepfake videos with over 1.5B views. The platform's rules, such as limiting targets to celebrities and prohibiting negative depictions, are poorly enforced, leading to the targeting of non-celebrities and the creation of abusive content. Deepfake creation is supported by a community that crowdsources facesets, utilizes open-source software like DeepFaceLab, and leverages cloud GPU providers, actively circumventing anti-abuse measures.
Approach
The authors systematically characterized the MrDeepFakes platform by crawling its public forum posts, video metadata, and user profiles. They performed quantitative measurements of forum activity and video attributes, combined with qualitative thematic analysis of forum threads, to understand marketplace economics, deepfake targets, and the deepfake creation process, including required software, hardware, and data.
Datasets
Public MrDeepFakes website content, including 42,986 video metadata entries, 8,198 forum threads comprising 43,350 posts, and public user profile information from 611,000 forum accounts and 595,000 tube site entries.
Model(s)
UNKNOWN
Author countries
United States