AI Forensics Report Finds Hugging Face Hosting Tools That Generate Nonconsensual Nude Deepfakes

A billion-dollar AI platform used by developers worldwide has a nonconsensual nude problem, according to a report published this week by AI Forensics, a European nonprofit research group.

The group tested nine of the most popular image-editing "Spaces" on Hugging Face, tools that let users run AI models directly through their browser. Seven of the nine could take a clothed photo of a woman and turn it into a topless one, no special hacking required.

AI Forensics didn't stop at testing the tools themselves. Researchers built their own decoy image-editing Spaces on the platform, engineered to accept prompts and images but never actually generate anything. Over one week, they logged more than 1,000 submissions.

The numbers are stark. AI Forensics says 73 percent of the prompts were sexual in nature. Of those, 83 percent were specifically asking to undress or sexualize the person in the photo. Ninety-five percent of the targets were women. And 6.7 percent of the sexual requests appeared to target children.

"Most of the Spaces [tested] can be used for generating nonconsensual intimate images, and users are actually using it for these purposes," said Paul Bouchaud, a lead researcher at AI Forensics. "This is not an empty threat, but actually people are using Hugging Face for that."

Wired conducted its own review of Hugging Face's site and, combined with findings from other researchers, found multiple pages actively promoting nudifying tools or models capable of generating sexualized images of named celebrities and politicians. That's a marketplace.

Hugging Face did not respond to Wired's questions about its content moderation practices or safety mechanisms. The company does have policies on the books banning child sexual abuse material and sexual deepfakes made "without explicit consent" or used for harassment. Some of the flagged pages disappeared after Wired reached out to the company. Whether that removal was a direct response to Wired's inquiry or unrelated moderation remains unclear, because Hugging Face didn't say anything.

The Bigger Pattern

This isn't an isolated platform problem. Nudifying apps and undress bots have exploded alongside generative AI's broader growth. The worst-known example: Elon Musk's Grok chatbot, which was used to produce millions of sexualized images of women and girls before controls were tightened.

Governments are starting to respond. US law enforcement has seized multiple deepfake hosting websites over the past year, according to Wired's reporting. The EU and UK have both drawn up plans to ban "nudify" apps outright by the end of 2026.

Hugging Face isn't a fringe operator. It's one of the largest open-source AI repositories on the planet, valued in the billions, and used daily by researchers, startups, and hobbyists to share and run AI models. That scale is exactly the problem. A platform that size hosting even a handful of nonconsensual nudify tools reaches an enormous number of potential victims.

What's Proven, What's Not

AI Forensics tested real tools on Hugging Face and got real nonconsensual results in seven of nine cases. The decoy-Space data on prompt behavior is AI Forensics' own methodology, not independently replicated by another outfit as of this reporting, though the volume and consistency of the findings—1,000-plus prompts over one week—make it hard to dismiss as a fluke.

What's not established is intent on Hugging Face's part. There's no evidence the company built these tools or wants this use case. The honest read is a moderation failure, not a deliberate business model. Open platforms that let anyone upload models face a genuine tension: they enable legitimate research and innovation, but the same openness lets bad actors upload garbage with minimal vetting. Critics of heavy-handed platform regulation have a fair point that banning entire categories of tools can also kill legitimate photo-editing and research use cases. That's a real tradeoff, not a strawman.

But a company sitting on billions in valuation, with named investors and enterprise customers, not responding to a single question from a major outlet about how it moderates content that includes apparent child-targeted sexual requests, is a defensible target for criticism regardless of political lean.

No US regulatory action against Hugging Face specifically has been announced. No charges have been filed. The EU and UK bans on nudify apps are still in the planning stage and haven't taken effect. Whether Hugging Face's quiet removals continue, or whether regulators eventually name the platform directly, remains an open question.