AI · 20h ago
Hugging Face hosts thousands of models that generate non-consensual deepfakes
Researchers found that most models on Hugging Face can create non-consensual deepfakes. The platform hosts over 2,000 models flagged for generating explicit content without consent. This raises serious safety and ethical concerns for the AI community.
Meridian48 take
Hugging Face's open model ecosystem lacks adequate safeguards, highlighting a systemic issue in AI content moderation.
Read the full reporting
Hugging Face reportedly plagued with AI models generating adult deepfakes →
Engadget
deepfakesai-safety