Hugging Face Spaces Enable Mass Production of Nonconsensual Deepfake Nudes
A new audit by AI Forensics found that most top image-editing apps on the platform can strip clothing from photos with simple prompts, and users are actively exploiting this capability.
Hugging Face has become a primary venue for generating nonconsensual sexual deepfakes, according to a report published this week by the European nonprofit AI Forensics. The findings expose a critical gap in safety infrastructure across the open-source AI ecosystem.
The Scale of Abuse
Researchers tested the nine most-popular image-editing Spaces on Hugging Face and found that seven could easily convert clothed photos of women into topless images using straightforward text prompts.
To measure actual usage patterns, AI Forensics deployed a honeypot account for one week. Of more than 1,000 prompts logged, 73 percent were sexual in nature. Among those sexual requests, 83 percent sought to undress or sexualize the subject, and 95 percent targeted women. Most alarmingly, 6.7 percent of sexual requests appeared to target children.
"Most of the Spaces can be used for generating nonconsensual intimate images, and users are actually using it for these purposes," said Paul Bouchaud, lead researcher at AI Forensics. "This is not an empty threat, but actually people are using Hugging Face for that."
Platform-Level Safeguards Missing
Unlike proprietary AI services from Google and OpenAI, which implement guardrails to block explicit imagery, Hugging Face leaves moderation to individual developers. Most do not add safeguards.
"No safeguards at all are being implemented at a platform level," Bouchaud said. "Only the developer can, if they want, implement some, and most of them do not."
Hugging Face's content policies officially prohibit child sexual abuse material and sexual deepfakes created without explicit consent. However, the report suggests these rules are not enforced through technical measures at the platform layer. WIRED reported that Hugging Face did not respond to numerous questions about its content moderation and safety practices.
Industrialization of Digital Abuse
The findings matter because Hugging Face serves as the primary distribution hub for open-source AI models. The lack of platform-level filtering enables what researchers describe as the "industrialization" of digital abuse, allowing users to create harmful nonconsensual intimate imagery with minimal effort.
The report underscores a broader tension in the AI industry: while open-source models promote accessibility and innovation, they also bypass the safety restrictions found in proprietary systems. As image-generation tools grow more powerful and widely available, the burden of preventing abuse increasingly falls on end users and downstream platforms rather than the infrastructure providers hosting these models.