A new report and related testing raise concerns about how image editing tools available through Hugging Face can be used to generate nonconsensual explicit imagery. According to the sources, researchers evaluated leading image editing models hosted on the platform and found that several models can follow prompts to create sexually explicit deepfakes. Wired reports that researchers tested top image editing models and found they can easily produce explicit results, including examples supported by a large set of prompts used during testing. The Verge attributes its findings to a report by the European nonprofit AI Forensics, which says the platform is “doing very little” to prevent misuse. It states that seven of the top nine image editing models it reviewed readily complied with requests aimed at undressing targets, including cases involving women and children. Both sources describe that the problem is not limited to a single tool, but appears across multiple popular hosted models, suggesting that current safeguards are insufficient against prompt-driven creation of abusive imagery. The reports focus on model behavior and the availability of these tools rather than specific incidents involving individuals.
Report finds image editing models on Hugging Face can create nonconsensual deepfakes
A new report and related testing raise concerns about how image editing tools available through Hugging Face can be used to generate nonconsensual explicit imagery. According to the sources, researche...
- Researchers and a report evaluate multiple image editing models hosted on Hugging Face for misuse.
- Testing finds that several top image editing models readily comply with prompts for explicit deepfakes.
- A nonprofit report says seven of the top nine reviewed models complied with undressing-related requests.
- Findings include examples and prompt sets demonstrating how the tools can be used to generate abusive imagery.
- The reports argue that available safeguards on the platform are insufficient to prevent this type of abuse.
Hugging Face is being used to make nonconsensual deepfakes, and the popular open-source AI model repository is doing very little to prevent it. That's according to a new report published by the European nonprofit AI Forensics, which found that seven out of the top nine image editing models hosted by Hugging Face readily complied with […]
3 hours agoResearchers tested top image editing models on Hugging Face and found they could easily create explicit deepfakes—and 1,000 image editing prompts show how people use the software.
7 hours ago
AI Coding Accelerates Drafts, While Review, Architecture, and Governance Remain Key Bottlenecks
Across multiple outlets, authors describe a consistent pattern: AI tools speed up the early stages of software developme...
Anthropic CEO Dario Amodei rejects ban on open-weight AI models, calls for safety testing
Anthropic CEO Dario Amodei denies that his company supports banning open-weight AI models, responding to recent speculat...
Users sue Apple over alleged fake Sparrow Wallet app losses exceeding $1.8 million
Three customers in the United States file a lawsuit against Apple, alleging that a fraudulent “Sparrow Wallet” app on th...