One 4Chan community is using a new technique to manipulate AI models into producing deepfake pornography of famous women, with one victim being as young as 19 years old.
A recent campaign by a community on the anonymous online imageboard platform 4Chan has revealed a new wave of attacks affecting young women.
4Chan users are exploiting female Olympic athletes by nudifying their images and spreading them across the internet.
During the 2026 Winter Olympics, users began generating and sharing non-consensual pornography of several US Olympic athletes.
The victims included gold medal-winning ice skater Alysa Liu, other US Olympians like Amber Glenn, Isabeau Levito, and Mikaela Shiffrin, as well as Eileen Gu, who competed for China.
American figure skater and Redbull athlete Isabeau Levito is only 19 years old.
During its investigation, Graphika identified that 4Chan users requested pornographic images altered by AI after having submitted clothed images of the 20-year-old US Olympic figure skater Alysa Liu.
4Chan users posted alleged images of 22-year-old freestyle skater Eileen Gu, which were blurred by Graphika.
Creating undressing agents is easier than we thought
The bad actors supposedly used low-ranked adaptations (LoRAs) to create this non-consensual deepfake imagery.
This method allows users to adapt a large language model to carry out a specific task without retraining the entire model.
“LoRA freezes the original weights and parameters of the model as they are. Then, on top of the original model, it adds a lightweight addition called a low-rank matrix, which is then applied to new inputs to get results specific to the context,” writes IBM.
Instead of retraining the model, which would take a lot of time, this method allows a user to keep its original components while adding new instructions, helping the model generate precise outputs tailored to the user's needs.
Have thoughts about this topic? Others do, too. Join them in the discussion.
While there are common use cases, such as creating fine-tuned GPTs, otherwise known as agents, to carry out specified tasks such as coding, Graphika has identified that LoRA is being used for nefarious purposes.
“To create a LoRA, users fine-tune a model using publicly available photos of a particular individual, effectively training the system to more accurately replicate their appearance,” the report claims.
Once these LoRA files have been created, the owner can share them with others, as they now contain the necessary adjustments.
From there, other users can “plug them into their own locally run models,” allowing more people to access the file and generate images targeted to specific individuals.
Deepfake pornography is an epidemic
We’ve seen this happen with various famous women, including singer Taylor Swift, whose deepfake image amassed over 47 million views on X.
Actor Scarlett Johansson has previously campaigned against the creation of deepfakes after her likeness was misappropriated countless times.
While users seem to be sharing this deepfake porn for free on 4Chan, criminals have earned tens of thousands of dollars by peddling non-consensual deepfake porn.
Tetsuro Chiba illegally distributed packages of AI-generated celebrity porn involving over 300 female celebrities.
The 31-year-old man had a collection totalling 520,000 images of women and allegedly made over $70,000 in profits.
Not only have we witnessed non-consensual deepfake porn on a smaller scale, but we’ve seen big players failing to stop the generation and proliferation of this kind of material.
Hordes of users used Elon Musk’s Grok to generate non-consensual explicit images of women and children.
The model was capable of creating this imagery, and Musk did little about it until absolutely necessary.
The problem with 4Chan
4Chan, the controversial anonymous online imageboard, has a pretty bad reputation.
The site has been the subject of various controversies due to the proliferation of misinformation on its platform.
Not only is 4Chan guilty of hosting conspiracy theories and misinformation on its site, but it’s also hosted illegal pornography like child sexual abuse material.
The UK telecommunications watchdog, Ofcom, launched an investigation into 4Chan in June 2025 over the proliferation of illegal content on its site, among other concerns.
“There are reasonable grounds for believing that the provider has contravened its duties under multiple sections of the Online Safety Act,” Ofcom said.
Unlock more exclusive Cybernews content on YouTube.
Your email address will not be published. Required fields are markedmarked