YouTube and X Are Sending Millions of Users to Nudify Apps, Report Says

YouTube and X have become major referral routes to so-called nudify apps, according to a new report that found mainstream social platforms drove more than 5.7 million visits to sites that generate nonconsensual explicit deepfakes between December 2025 and March 2026. The findings matter because they suggest some of the internet’s biggest platforms are not just hosting the problem of nonconsensual intimate imagery, but actively helping users find the tools used to create it.

The report, published by the Institute for Strategic Dialogue, an anti-extremism and anti-disinformation group, examined the top apps and websites used to create nonconsensual explicit synthetic images and mapped how people discovered them online. Its central conclusion is stark: a surprising share of traffic is coming from household-name platforms that say they prohibit sexually explicit content and links to such material.

That disconnect is especially alarming as lawmakers and regulators around the world try to catch up with rapidly spreading AI tools that can strip clothing from photos, generate fake nudity, and weaponize images against victims. The study arrives as researchers, advocates, and journalists have documented a steady rise in deepfake abuse in schools, workplaces, and personal relationships.

What the report found

The ISD analysis focused on the online ecosystem around nudify services, which are websites and apps that use AI to create explicit images of people without their consent. These tools are often marketed with euphemisms, but their use is frequently tied to harassment, humiliation, and extortion.

Across the four-month period studied, social networks sent more than 5.7 million visits to nudify sites. YouTube was the largest source by a wide margin, accounting for 1.82 million visits, or more than 30 percent of the referrals identified in the report. X followed with more than 1.3 million visits.

The researchers said much of the traffic appears to have been driven by posts and videos optimized around searches such as “undress app” and “nudify app.” In many cases, the content did not merely discuss the tools in passing. It reviewed specific apps, promoted them, offered discount codes, and pointed users directly to off-platform services.

According to the report’s authors, this pattern is in direct tension with YouTube’s own rules, which bar sexually explicit content and external links to material of that kind.

The study’s wording is notable because it argues that material linked to nudify services should fall squarely within those policies. Yet, the report says, the relevant posts remained easy to find and use, effectively turning YouTube into a gateway to websites that help people create nonconsensual explicit imagery.

Why YouTube and X matter so much

YouTube and X matter because they are not fringe forums. They are large, algorithmically driven discovery engines with enormous reach, and that makes them powerful distribution systems for harmful content even when the platforms do not host the final product themselves.

In practice, the report suggests, the referral problem can be as damaging as direct hosting. A user searching for a tool to fake nudity does not need to stumble across a dark-web marketplace or a niche message board. They may instead find a polished video tutorial on a mainstream platform, complete with links, comments, and promotional copy that make the service feel normalized.

Melanie Smith, the senior director of research and policy at ISD, said the study showed more than passive referral traffic. In her view, some platform content was helping users understand how to use these tools, not just pointing them toward them.

Smith said that in a number of cases the platform was facilitating use of the services rather than simply appearing in the referral logs. She also noted that policies banning links to sexually explicit websites should, in theory, apply to leaked nudes, revenge porn, and similar abuse, but that enforcement does not appear to be comprehensive.

That enforcement gap is at the heart of the story. Platforms often have formal rules against explicit material, sexual exploitation, and nonconsensual imagery. The challenge, the report suggests, is not simply what those rules say, but how consistently they are enforced across different kinds of content, creators, languages, and formats.

How nudify apps spread online

Nudify apps spread through a mix of search optimization, social video, reposting, affiliate marketing, and user-generated promotion. That mix is important because it means these services are often discovered through ordinary browsing behavior rather than through obscure or technically difficult channels.

Search terms, tutorials and promo codes

One of the most troubling findings was the role of search behavior. People looking for explicit deepfake tools appear to encounter videos and posts that are built around obvious keywords, making them easier to surface through platform search and recommendation systems.

Some of the content reviewed by ISD was straightforward promotion. Other examples included tutorials, demonstrations, and videos that offered free credits or discount codes. That kind of promotion lowers the barrier to entry and helps turn an abusive service into something that looks like a routine consumer app.

Why low prices do not mean low harm

Some of these services cost as little as $1 per image, making them cheap enough to be used impulsively, repeatedly, and at scale. But the report warns that affordability does not mean the underlying business is small. In fact, these tools can still be highly profitable because the volume of users and images can be large.

A separate WIRED investigation previously estimated that the industry could be bringing in as much as $36 million a year in combined revenue. That scale helps explain why the services continue to proliferate even as public scrutiny intensifies.

What kinds of abuse are these tools being used for?

These tools are being used for harassment, coercion, humiliation, and in some cases employment sabotage. While many people assume the motivation behind nudify apps is sexual gratification, the ISD report found a more varied and disturbing set of use cases.

Common targets include current and former romantic partners, but the research also noted examples involving family members such as sisters and cousins. That pattern highlights how easily these tools can be folded into interpersonal abuse and domestic conflict.

Smith said the requests observed by researchers were not always sexual in nature. In some cases, users appeared to be trying to get someone fired, discredit a person socially, or compromise their livelihood. That broadens the picture from explicit content abuse to a wider category of malicious digital harm.

Beyond embarrassment

The real-world consequences can be severe. Victims of synthetic sexual abuse may face reputational damage, emotional trauma, school discipline, workplace fallout, blackmail, and online pile-ons. Because the material can be generated quickly and shared widely, the harm can spread faster than victims can respond.

That is one reason why advocates have pushed for stronger platform intervention and faster takedown mechanisms. Once a synthetic image is made public, copying and reuploading it can make removal much harder.

How are platforms responding?

Platforms say they have rules, but the report suggests enforcement remains uneven. In response to WIRED, YouTube said it has strict policies against unwanted sexualization, including nonconsensually shared intimate imagery, and that those rules apply to both on-platform content and external links. The company also said its policies cover altered or synthetic material that realistically simulates nudity.

X did not respond to a request for comment. The company has previously faced backlash over users exploiting its AI chatbot Grok to create nude or sexualized images of women without their consent, including some minors. After that controversy, X said it would limit access to Grok to paying users and reiterated that it had zero tolerance for child sexual exploitation, nonconsensual nudity, and unwanted sexual content.

The problem, as the ISD report frames it, is not whether the rules exist on paper. It is whether moderation systems, recommendation systems, reporting processes, and enforcement decisions are strong enough to stop those rules from being bypassed at scale.

Policy versus enforcement

There is a recurring pattern in online safety debates: a company announces a policy, publishes a transparency update, and then struggles to police the volume and creativity of user behavior. Nudify promotion appears to be another example of that gap.

Because the content may be framed as a review, a tutorial, a meme, or a “tool demo,” it can evade simple keyword-based moderation. That means effective enforcement often requires contextual judgment, better detection of abusive intent, and faster removal of outbound links that facilitate harm.

What the law says now

NCII, or nonconsensual intimate imagery, is illegal in the United States. Federal rules now require major social platforms to remove such material within 48 hours after a victim submits a valid takedown request under the Take It Down Act, which fully took effect in May.

That legal framework is designed to make response times faster and victim-centered. It also reflects a broader recognition that online sexual abuse needs more than ad hoc moderation; it needs a formal obligation to act.

At the state level, anti-deepfake and anti-NCII laws have expanded rapidly. Minnesota became the first state to specifically ban nudification apps in May 2026, marking a rare example of legislation aimed directly at the tools rather than only at the images they produce.

Even so, the report argues that the spread of nudify services has not slowed. If anything, they are becoming easier to access, cheaper to use, and simpler to promote.

Why legislation alone is not enough

Law can set standards, but it often moves far more slowly than the platforms and tools it seeks to regulate. That mismatch leaves a large window in which abuse can continue, especially when services operate across borders and platforms with different enforcement systems.

One difficulty is jurisdiction. A website can be hosted in one country, marketed in another, and used by someone halfway around the world. Even if one market tightens restrictions, the service may still remain reachable elsewhere through mirror sites, clones, or new domains.

Another challenge is scale. Social platforms host billions of posts, videos, comments, and links. Moderators and automated systems must triage huge amounts of content, and bad actors adapt quickly when a loophole closes.

What should change?

The ISD report calls for a coordinated response that includes online enforcement, offline support, policy interventions, and digital literacy efforts. In practical terms, that means more than just deleting a few bad accounts.

• Tighter moderation of search-based promotion for abusive tools
• Stronger enforcement of rules against outbound links to explicit services
• Clearer victim reporting and rapid takedown pathways
• School-based digital literacy and safety education
• Cross-platform coordination on repeat offenders and affiliate networks

Education matters because many victims first encounter this threat in schools, where social pressure, smartphones, and peer conflict can create a volatile environment. Recent reporting has identified deepfake abuse in more than 90 schools worldwide, underscoring that this is no longer a niche internet crime.

How did a niche abuse tool become mainstream?

It became mainstream because the barriers to creation dropped. Generative AI, cheap cloud services, and easy-to-use interfaces have transformed what once required technical skill into something anyone can attempt in minutes.

That shift does not merely increase the number of bad actors. It changes the social meaning of the act itself. When a harmful tool is marketed like a novelty app, it can seem less like digital abuse and more like a prank or personal experiment.

Social platforms intensify that dynamic. Recommendation systems reward engagement, and sensational content tends to attract clicks. A video promising a shocking or illicit result can travel quickly, especially if it is framed as entertainment, curiosity, or consumer advice.

Normalization is part of the danger

Normalization matters because once harmful behavior becomes familiar, users are less likely to see it as criminal or abusive. The report suggests nudify promotion has reached that stage in some corners of the mainstream internet, where it can be presented as a clever shortcut rather than a violation of consent.

That is exactly why researchers and advocates see the issue as more than a content moderation headache. They view it as a consent crisis, a privacy crisis, and a public safety issue all at once.

Timeline of the latest escalation

The recent controversy did not emerge in isolation. It is part of a broader escalation across products, platforms, and laws over the past year.

What happens next?

The likely next step is more scrutiny of platform search, link-sharing, and monetization pathways. If the report is correct, the issue is not just a handful of posts violating the rules, but a recurring ecosystem that makes abusive tools easy to discover.

For YouTube and X, that means the pressure will probably focus on detection and enforcement rather than statements of principle. Regulators and researchers are increasingly asking whether companies can prove that their systems are actually suppressing harmful referral networks, not simply reacting after the fact.

For victims, the stakes are immediate. A single synthetic image can trigger harassment, reputational harm, and long-term emotional distress. For schools, parents, and employers, the report is another warning that the line between internet prank and real-world abuse has all but disappeared.

And for the broader tech industry, the message is hard to miss: if mainstream platforms are helping users find tools that generate nonconsensual explicit images, then content moderation cannot be treated as a side issue. It is central to whether the next wave of AI products becomes a useful technology or a scalable abuse machine.

In that sense, the ISD report is not only about one class of websites. It is about the infrastructure of harm around them — the platforms, search behavior, incentives, and policy gaps that let abuse spread faster than the systems designed to stop it.