Hugging Face reportedly plagued with AI models generating adult deepfakes

The 30-second read

Recent investigations into Hugging Face reveal that most of its hosted AI models are capable of generating non-consensual deepfakes. Corporate security leads should evaluate potential brand and liability risks associated with open-source model repositories currently lacking robust content safeguards.

• 01Researchers found a majority of Hugging Face models generate unauthorized adult deepfake content
• 02Findings highlight a widespread lack of effective safety filters across popular open-source AI platforms
• 03Corporate legal and security teams must assess exposure to non-consensual content generation tools
• 04Platform integrity concerns may impact enterprise adoption of unfiltered open-source machine learning models

AI-generated summary · Verify at source