How a 4chan Community Uses AI to Create Fake Intimate Images — and Why It Matters

New research from the Institute for Strategic Dialogue shows that a group on 4chan is using artificial intelligence to create fake explicit photographs of real women without their consent. The research found that 4chan's adult requests board serves as a hub where people ask others to create these images, and then those images spread to other platforms like Telegram and Discord.

The study looked at thousands of posts from December 2025 through March 2026. Here's how it works: someone on 4chan posts a photograph of a woman and asks the community to create a fake explicit version of it. People the researchers call "wizards" — because they have the technical skills to use the AI tools — then create the image and post it back.

How the Community Works

The research shows that this isn't just about individual people making images in secret. Instead, it's a group activity where people who create the images get praised by others in the community. This praise and recognition encourages more people to participate. Essentially, the community is bonding over targeting women.

Most of the people being targeted are women, which fits a larger pattern: researchers have found that 96 percent of deepfake videos online are explicit images made without anyone's permission.

The images created on 4chan don't stay there. People copy them and share them on messaging apps and other websites. This makes the problem bigger because the fake images spread far beyond where they started.

What Technology Makes This Possible

The AI technology that creates these fake images works by taking a real photograph and swapping in a person's face onto existing explicit video or photos — kind of like a very sophisticated version of the face-swap filters you might use on social media, but applied to explicit material.

What makes this concerning is that the tools have become easier to use. A few years ago, you needed serious technical training to create these images. Now, the software is designed so that almost anyone can use it, even without deep knowledge of artificial intelligence.

What Governments and Platforms Are Doing

Several places have started taking action. Britain recently made it illegal to create these fake images. Sweden has similar laws. In the United States, the federal government has arrested people under a law called the TAKE IT DOWN Act, and the Department of Justice has prosecuted people for making both adult and child-targeted synthetic imagery.

One recent case shows how this is evolving: the city of Baltimore is suing Elon Musk's AI company xAI because their Grok chatbot can be used to generate these fake images.

The FBI has also warned the public that criminals use these fake images for blackmail schemes, where they threaten to share the images unless the victim pays money.

Platforms have tried to fight back. When major websites that hosted these images were shut down, the amount of new content dropped. But the research suggests that the creators have adapted and found new ways to operate.

What This Reveals About Online Harassment

Researchers have been studying how AI tools get used as weapons against women online. They point out that this fits a longer pattern: over the years, people have used whatever technology is available to harass others — from coordinated bullying campaigns to doxxing (publishing someone's private information) to swatting (sending emergency responders to someone's home as a prank).

Each time a new technology emerges, the people doing the harassing figure out how to use it. The 4chan request system is the latest version: instead of one person doing the harassing alone, it's now a group activity that builds community through shared participation in targeting someone.

This is worth paying attention to. What makes this version different from earlier forms of online harassment is that it requires the target to actually exist. The AI needs a real photograph of a real person. That means anyone can be targeted — a neighbor, a colleague, someone from high school.

Why This Is Hard to Stop

The researchers note that traditional approaches to stopping this kind of content don't work very well anymore. In the past, platforms could remove bad images once they found them. But this situation is trickier because the problem isn't just the final image — it's the entire system around it: the requests, the praise for creators, the sharing networks across multiple platforms.

To actually stop this, platforms and law enforcement would need to disrupt the community itself, not just delete individual images. That's a much harder problem to solve.

The research from the Institute for Strategic Dialogue documents exactly how this ecosystem works, which gives policymakers and technology companies a clearer picture of what they're dealing with.