A comprehensive investigation by the Institute for Strategic Dialogue (ISD) has revealed a troubling paradox in the digital landscape: while major social media platforms maintain public-facing policies against sexually explicit content, they are simultaneously serving as the primary referral engines for "nudify" applications. These tools, which utilize generative artificial intelligence to digitally remove clothing from images of individuals without their consent, have seen a massive surge in popularity, fueled largely by traffic from mainstream sites such as YouTube and X (formerly Twitter). The ISD report, titled "Mapping the Nudify Tools Ecosystem," details how these platforms have transitioned from passive hosts to active gateways, effectively connecting millions of users to services that facilitate the creation of nonconsensual intimate images (NCII).
The Rise of the Nudification Ecosystem
The "nudify" phenomenon represents one of the most pervasive and harmful applications of generative AI technology. Unlike early deepfake efforts that required significant technical expertise and high-end hardware, modern nudification tools are often browser-based or available as simple mobile applications. These services allow users to upload a photo of any individual—be it a celebrity, a colleague, or a private acquaintance—and generate a realistically rendered nude version of that person in seconds.
The ISD study focused on the top 10 apps and websites dedicated to this practice, analyzing how users discover these services. Contrary to the assumption that such tools are relegated to the "dark web" or unregulated fringe forums like 4chan and 8kun, the data suggests that the vast majority of users are directed to these sites via standard search queries and social media recommendations. Between December 2025 and March 2026, social networks were responsible for driving more than 5.7 million visits to nudify websites. This high volume of traffic highlights a systemic failure in the moderation algorithms and policy enforcement mechanisms of the world’s largest tech companies.
Chronology of AI-Generated Abuse
The evolution of nonconsensual deepfake technology has moved at a rapid pace over the last decade, leading to the current crisis identified by the ISD.
- 2017–2018: The emergence of "Deepfakes" on Reddit, primarily used to swap celebrity faces onto adult film stars. This required specialized software and significant computing power.
- 2019: The launch of "DeepNude," the first widely publicized "nudify" application. While the creator took the site down shortly after launch due to public outcry, the underlying code was mirrored and distributed across the internet.
- 2022–2023: The "Generative AI Boom." The release of open-source models like Stable Diffusion allowed developers to create sophisticated "undressing" algorithms that could be hosted on standard web servers.
- 2024–2025: The commercialization phase. Nudify services began adopting "Software as a Service" (SaaS) models, offering tiered subscriptions, affiliate programs, and aggressive marketing campaigns on mainstream platforms.
- 2026 (Present): The ISD report confirms that these tools have reached a level of mainstream accessibility where social media platforms act as the primary marketing funnel, bypassing traditional gatekeeping efforts.
Analyzing the Data: YouTube and X as Primary Drivers
The ISD report provides a granular look at where the 5.7 million referrals originated. YouTube emerged as the single largest contributor to the ecosystem, accounting for 1.82 million site visits, or roughly 32 percent of the total referral traffic. The mechanism for this traffic is often instructional or promotional. Users searching for terms like "undress app" or "nudify app" are presented with videos that review specific tools, provide tutorials on how to achieve the most "realistic" results, and include direct links to the services in the video descriptions.
Furthermore, many of these YouTube videos include promotional codes that offer users free "credits" to generate images. This creates a direct financial incentive for content creators to promote these tools, as they often receive commissions through affiliate marketing schemes.
X (formerly Twitter) followed closely as the second-largest driver, responsible for over 1.3 million visits. On X, the proliferation is often driven by automated bots or "shill" accounts that reply to trending posts with links to nudify services, often accompanied by "before and after" samples that bypass automated filters by using clever cropping or slight blurring that human eyes can easily interpret but algorithms may miss.
The Disconnect Between Policy and Enforcement
The findings highlight a significant gap between the stated policies of tech giants and the reality of their platforms’ operations. YouTube’s official guidelines strictly prohibit sexually explicit content, a category that logically extends to tools designed to create such imagery.
Melanie Smith, the senior director of research and policy for ISD, noted that YouTube’s role is not merely passive. "In a lot of these cases, it was facilitating the use of these tools as well," Smith told WIRED. She emphasized that while YouTube’s policies explicitly forbid linking to or advertising sexually explicit websites—similar to how they treat links to OnlyFans or pornography hubs—enforcement remains inconsistent.
In a statement responding to the findings, YouTube spokesperson Boot Bullwinkle reiterated the company’s stance: "We have strict policies prohibiting content that includes unwanted sexualization, such as nonconsensually shared intimate imagery." Bullwinkle added that these policies apply to both on-platform content and external links, including "altered or synthetic content that realistically simulates nudity." However, the ISD report suggests that despite these policies, the sheer volume of content and the speed at which new links are generated have overwhelmed YouTube’s moderation infrastructure.
The Economics of Digital Violation
The nudify industry has become a highly lucrative sector within the broader AI market. The ISD report, building on previous investigative work, suggests that these platforms are collectively generating as much as $36 million in annual revenue.
The pricing models are designed for low-friction entry. Many sites allow users to generate a single image for as little as $1, often using a "credit" system. Users purchase bundles of credits—for example, $10 for 12 images—which obfuscates the direct cost and encourages repeat usage. The low cost of entry, combined with the ease of use, has democratized the ability to commit digital harassment, moving it from the realm of hackers to the general public.
Target Demographics and Malicious Intent
One of the most disturbing findings of the ISD study concerns the targets of these applications and the motivations of the users. While celebrity targets remain common, there is a growing trend of "civilian" targeting. The study found that users frequently use these tools on:
- Current and former romantic partners (often as a form of "revenge porn").
- Colleagues and acquaintances.
- Relatives, including sisters and cousins.
Perhaps most alarming is the shift in motivation. While sexual gratification is a primary driver, a significant portion of users are motivated by a desire to inflict professional and social harm. Melanie Smith noted that many requests within the communities associated with these tools were focused on "getting people fired from jobs and compromising their livelihoods." By creating and disseminating these images to an individual’s employer or social circle, harassers can inflict devastating real-world consequences without ever physically touching the victim.
Broader Implications and the Path Forward
The findings of the ISD report have sparked renewed calls for legislative and regulatory action. Currently, the legal landscape regarding AI-generated NCII is a patchwork of inconsistent laws. In the United States, several states have passed laws specifically targeting deepfake pornography, and federal legislation like the DEFIANCE Act has been proposed to provide victims with a civil cause of action. However, the international nature of these websites—often hosted in jurisdictions with lax digital regulations—makes total eradication difficult.
The responsibility, according to digital rights advocates, lies heavily with the platforms that provide the audience. The ISD report suggests several areas for immediate improvement:
- Keyword Blacklisting: More aggressive filtering of search terms related to nudification.
- Affiliate Link Detection: Enhanced AI tools to identify and remove links that lead to known nudify domains, even when obscured by URL shorteners.
- Demonetization and Banning: Immediate removal of channels and accounts that provide tutorials or promotional codes for these services.
- Victim Support: Streamlined reporting processes for individuals who find their likenesses being used in promotional material for these apps.
The "nudify" ecosystem represents a critical failure in the safety-by-design principles of both generative AI and social media platforms. As the technology becomes more sophisticated and the images more indistinguishable from reality, the potential for lifelong trauma and professional ruin for victims grows exponentially. The ISD report serves as a stark reminder that as long as mainstream platforms continue to act as gateways, the industry of digital violation will continue to thrive at the expense of human dignity and safety.
