Ethical Hacking News
Researchers have found that Hugging Face's open-source AI platform is being exploited to create non-consensual deepfakes, raising concerns about digital consent and the need for greater regulation in the development and deployment of image generation models.
Hugging Face's image editing models have been found to easily create non-consensual deepfakes, including undressing a person in images. A recent report by AI Forensics found that 73% of prompts on Hugging Face were sexual in nature, with 95% targeting women. 6.7% of the sexual requests targeted apparent children. The platform has been hosting over 5,000 AI image models capable of creating non-consensual pornography. Image generation models often lack safety mechanisms to prevent the creation of undressed images without consent. The issue highlights the need for greater regulation and oversight in the development and deployment of image generation models.
Hugging Face, a popular open-source AI platform, has been under scrutiny lately due to its widespread problem with non-consensual deepfakes. A recent report by the European nonprofit AI Forensics found that researchers tested nine of the top image editing models on Hugging Face and discovered that seven of them easily changed a clothed image of a woman into a topless one.
The research also involved creating honey-pot-style image editing Spaces on Hugging Face, which were designed not to produce any images. However, AI Forensics researchers tracked over 1,000 prompts and images they received over a week, revealing that 73 percent of the prompts were sexual in nature. Among these, 83 percent were seeking to undress or sexualize the person who had submitted a photo, with 95 percent of these being women.
Moreover, AI Forensics researchers found that 6.7 percent of the sexual requests targeted apparent children. The report highlights how users are exploiting Hugging Face's image editing models for non-consensual intimate images, often without any regard for digital consent.
The AI Forensics' findings were corroborated by another review of materials on Hugging Face's website, which revealed multiple pages promoting nudifying technologies or AI models that could potentially create sexualized images of named celebrities and politicians. Some of these pages have been removed after WIRED contacted the company, but it is unclear if the removals are directly related to the nudifying services.
The issue with non-consensual deepfakes on Hugging Face is not an isolated incident. The platform has been hosting around 5,000 AI image models that can create images of real people, which have previously been used to create non-consensual pornography. Last month, Transformer reported that Hugging Face was hosting more than a dozen tools that can be used to generate sexual deepfakes of prominent political figures.
Other researchers, such as Leonie Oehmig and Benjamin Shultz, have also pointed out the risks associated with image generation models and face-swapping apps. These apps often possess capabilities to create undressed images of people without safety mechanisms, which can lead to non-consensual intimate content.
Furthermore, the AI Forensics' honey-pot provided a telling glimpse into how image generation models may be used to create harmful images of people without their consent. The researchers found that prompts published by them showed multiple requests for images to be edited to include the depiction of semen on women, changing their appearance to be using sex toys or performing other sexual acts.
The study highlights the need for greater regulation and oversight in the development and deployment of image generation models. While many mainstream generative AI models use safety mechanisms to try to prohibit the creation of undress-style images, the models inspected by AI Forensics appeared not to have any such safeguards.
"The most part of them [image models] are being used for generating non-consensual intimate images," says Paul Bouchaud, a lead researcher at AI Forensics. "Most of the spaces can be used for this purpose."
The issue is further complicated by the fact that many image generation models have been trained using sexual images from the internet and can create explicit content unless they deploy safety mechanisms.
In response to the growing concerns surrounding Hugging Face's deepfake nudes problem, it remains to be seen whether the company will take concrete steps to address this issue. However, the report by AI Forensics serves as a wake-up call for the tech industry to prioritize digital consent and ensure that image generation models are developed with safeguards against non-consensual content.
Related Information:
https://www.ethicalhackingnews.com/articles/Hugging-Faces-Deepfake-Nudes-Problem-A-Growing-Concern-for-Digital-Consent-ehn.shtml
https://www.wired.com/story/hugging-face-has-a-nonconsensual-deepfakes-problem/
https://dnyuz.com/2026/07/28/hugging-face-has-a-deepfake-nudes-problem/
https://savedelete.com/news/hugging-face-deepfake-nudes/
Published: Tue Jul 28 01:00:23 2026 by llama3.2 3B Q4_K_M
