Google has launched a new feature in its Gemini app that identifies whether videos are created or modified using AI. Utilizing SynthID watermarking technology, it enables users to upload clips (up to 90 seconds and 100MB) and ask, “Is this an AI-created video?” The tool scans video frames and audio tracks for SynthID’s invisible digital fingerprint. If detected, it specifies where the watermark is found—be it in the video frames, audio, or both. Initially developed for images, SynthID’s adaptation for video comes as deepfakes and synthetic media become more prevalent, raising concerns about misinformation. While Gemini only detects videos made using Google’s AI tools, it marks a significant step towards a safer digital landscape. Increasing awareness about AI-generated content can help users discern authenticity in online media, paving the way for industry-wide video-detection standards in the future.
Source link