Hugging Face Has a Deepfake Nudes Problem - WIRED

What the report says

Wired reported that the open-source AI platform Hugging Face is facing scrutiny after a new report from the European nonprofit AI Forensics found that widely accessible image-editing tools on the site could be used to generate nonconsensual sexual deepfakes. According to Wired, the researchers tested nine highly visible Hugging Face “Spaces,” which let users run AI models directly on the platform, and found that seven could turn an image of a clothed woman into a topless image using a simple prompt.

AI Forensics also created test Spaces that did not generate images but collected user inputs over one week. Wired reported that the group received more than 1,000 prompts and images, with 73% of prompts described as sexual. Of those sexual prompts, 83% sought to undress or sexualize the person in an uploaded image, 95% of the targets were women, and 6.7% appeared to involve children. Wired said its own review found Hugging Face pages promoting nudifying tools or models that could potentially sexualize named public figures.

Hugging Face did not respond to Wired’s questions before publication, but later provided a document saying the findings pointed to a gap in safety measures adopted by some developers and that the company was working on the issue. The company also disputed aspects of AI Forensics’ methodology and said the collected prompts were not representative of overall platform use.

The report comes as governments and law enforcement move against deepfake abuse, including proposed bans on “nudify” apps in the EU and UK. Wired noted that Hugging Face policies prohibit child sexual abuse material and nonconsensual sexual deepfakes, but researchers argued platform-level safeguards remain inadequate.