Meta Unveils Open-Source Tool 'Video Seal' to Combat AI Deepfakes

The tool applies imperceptible watermarks to AI-generated videos, aiming to curb misinformation and enhance content authenticity.

Meta's new Video Seal tool embeds invisible watermarks on AI-generated videos to help identify synthetic content and prevent misuse.
The tool is open source and designed for integration into existing software, with claims of resilience against compression, blurring, and cropping.
Video Seal joins Meta's existing watermarking tools, Audio Seal and Watermark Anything, targeting video, audio, and image-based AI content respectively.
Meta plans to encourage adoption by launching a public leaderboard, Meta Omni Seal Bench, to compare watermarking methods and hosting workshops at AI conferences.
The release comes as deepfake-related fraud has surged globally, raising concerns about misinformation, fraud, and the difficulty of distinguishing real from fake content.