Most of us think the big tech companies have walls. High ones. They keep out the nasty stuff—nonconsensual intimate images, sexual explicit content. The rules are written. Signed. Enforced? That part is up for debate.
A new report from the Institute for Strategic Dialog (ISD) says those walls have gates. Open ones.
It’s not some fringe forum hiding in the dark corners of the web. It’s YouTube. It’s X. These mainstream giants are actively driving traffic to sites that digitally undress people against their will.
The Numbers Don’t Lie
The study looks at the top tools used for making nonconsensual explicit deepfates. Then it looks at where users find them. You’d expect the referral traffic to come from unregulated dumps like 4chan. It doesn’t.
Between December 2025 and early 2026, social networks sent 5.7 million people to these sites.
YouTube led the charge. Responsible for 1.8 million visits alone. That is over 30 percent of all the referrals.
How? Search results. Someone types “undress app” into YouTube. Up pops a review. Or a tutorial. Or a promo code for free credits on a nudification tool. It works.
X followed close behind. 1.3 million visits.
“It wasn’t just that YouTube was passive.”
Melanie Smith from ISD tells WIRED that the platforms weren’t just hosting links. They were facilitating the use.
YouTube’s own policy prohibits sexually explicit content. It also bans links to such sites. In theory, a nudification tool generator violates both. In practice? Smith says it seems the rules exist on paper, not on the screen.
A spokesperson for YouTube, Boot Bullwinkle, pushed back. He told WIRED that strict policies are in place against “unwanted sexualization” and synthetic nudity. The policy applies to internal content and external links. The implication is clear. The system is working, he says.
The evidence suggests otherwise.
Why Do It?
Cost is low. You can generate a deepfake nude for $1.
That might not sound like much. Until you look at the industry as a whole. These tools raked in an estimated $36 million last year.
Who is being targeted? Ex-girlfriends. Current partners.
Sisters. Cousins.
Smith found the motives are rarely purely sexual. Often it’s spite. People want to get targets fired. They want to ruin livelihoods. It’s a weapon. And the trigger is cheap.
X’s Own Chatbot Joined the Club
The problem isn’t just links to external sites. Sometimes the platform provides the tool.
In January 2024, users turned Grok—X’s AI chatbot—against women. They asked it to generate nude images of strangers. Minors included. The backlash was immediate and severe.
X responded by limiting access. Only paid users can now talk to Grok.
“We remain committed to making X safe,” a company statement read.
Zero tolerance for exploitation. The words feel heavy. Did nothing happen? Not necessarily. But the genie is out of the bottle. X declined to comment further when asked for updates.
Laws Lag Behind Tech
Nonconsensual imagery is illegal in the US. The Take It Down Act requires platforms to remove content within 48 hours of a takedown request. Minnesota went further, becoming the first state to outright ban nudification apps in May 2026.
Laws matter. But they are slow.
The apps are faster. Smoother. Cheaper.
ISD calls for coordinated action. Regulation. School workshops. Digital literacy. It sounds sensible. Logical.
A WIRED investigation found deepfake incidents in over 90 schools worldwide. The trend isn’t slowing. If anything, the infrastructure for harassment is getting more efficient every month.
The gate is open. The platforms are pointing the way. And nobody seems to know how to close it.
