Tag: deepfakes

Showing 1-3 of 3

Aug 08, 20267 min readModel Release BriefIntermediate

Community-hosted image models on Hugging Face accept 'undress' prompts, enabling nonconsensual sexual deepfakes

AI Forensics found seven of nine top Hugging Face image-edit models comply with simple "undress" prompts, enabling realistic nonconsensual sexual deepfakes - steps teams should take.

Hugging Face deepfakes content-moderation model-hosting safety

+2 more

AI policy image-editing

NewsUnited Kingdom

Open

Feb 05, 20267 min readModel Release BriefIntermediate5 min build

Provenance labels and metadata are failing as deepfakes scale

The Verge argues provenance manifests and in-band labels are brittle: transcoding, resharing, and model realism are breaking metadata-based safeguards against manipulated images and video.

deepfakes provenance c2pa metadata disinformation

+3 more

generative-ai policy journalism

NewsFrance

Open

Jan 30, 20267 min readTooling Deep DiveIntermediate5 min build

Civitai LoRA files and bounties enable bespoke deepfakes targeting real women

Stanford/Indiana research shows Civitai’s LoRA files and 'bounties' let users produce bespoke deepfakes—86% using LoRAs and 90% of requests targeted women.

civitai deepfakes lora marketplace moderation

+3 more

safety privacy a16z