Deepfakes Without Guardrails: The Unchecked Rise of Nonconsensual AI Generated Content A recent report from European nonprofit AI Forensics has exposed a disturbing trend: the increasing use of popular open source AI model repository Hugging Face to create nonconsensual deepfakes.
What's striking about this finding is not just the prevalence of these AI generated images, but also the lack of response from Hugging Face.
The report highlights seven out of nine top image editing models hosted by Hugging Face that can be easily coaxed into undressing women using simple prompts. This issue extends far beyond a single platform, however.