YouTube, X Funnel Millions to AI Deepfake Nudify Apps: Study
A new report reveals that major social media platforms like YouTube and X (formerly Twitter) are inadvertently driving millions of visits to websites that use generative AI to create non-consensual explicit deepfakes, commonly known as 'nudify' or 'undressing' apps. The Institute for Strategic Dialogue (ISD) published the study on July 13, 2025, examining the top 10 apps and websites used to produce such imagery and how users discover them.
According to the ISD, social networks generated over 5.7 million visits to nudify sites between December 2025 and March 2026. YouTube accounted for more than 30% of this referral traffic, approximately 1.82 million visits. X was the second-largest source, contributing over 1.3 million visits. Researchers found that these platforms hosted videos and posts that not only reviewed and promoted specific nudify apps but also provided links to promo codes for free credits.
Nudify apps use generative AI to realistically depict individuals without clothing in images or videos, often without the subject's consent. The report highlights that access to these tools is remarkably cheap, with some apps charging as low as $1 (roughly Rs 96) per image. Despite the low cost, the industry is highly profitable: a separate report by Wired estimated that some nudify apps generate as much as $36 million in collective revenue annually.
The findings challenge the assumption that such apps spread only through fringe or lightly moderated platforms like 4chan. Instead, the study shows that a large share of traffic originates from mainstream social media sites. The ISD argues that the guardrails put in place by platforms to prevent the dissemination of non-consensual intimate imagery (NCII) are inadequate. For instance, YouTube's policies prohibit sexually explicit content and links to such material, yet content violating these rules was easily discoverable.
Melanie Smith, the senior director of research and policy at ISD, told Wired: "It wasn't just that YouTube was a passive source of referral traffic. In a lot of these cases, it was facilitating the use of these tools as well." She noted that while YouTube's policies theoretically cover non-consensual imagery and revenge porn, enforcement appears insufficient.
In response to the study, YouTube spokesperson Boot Bullwinkle stated that the company has "strict policies prohibiting content that includes unwanted sexualization, such as non-consensually shared intimate imagery." He added that these policies apply to both content on YouTube and external links, including "altered or synthetic content that realistically simulates nudity." The report, however, suggests that enforcement gaps persist.
The ISD study underscores the urgent need for stronger moderation and enforcement to prevent mainstream platforms from becoming gateways to harmful AI-generated content.
