Link: Google’s Gemini app can check videos to see if they were made with Google AI

Google has enhanced its Gemini app by introducing an AI verification tool for videos. This feature determines whether a video is AI-generated using Google's technology by recognizing a unique watermark called SynthID.

Upon receiving a user query, Gemini analyzes both the visual and audio components of the video to find instances of this watermark. Users are then informed of the specific times where the watermark is detected, rather than just receiving a simple yes or no answer.

The tool was first applied to images in November, offering verification for visuals altered or created by Google's AI. Although Gemini can now authenticate videos up to 90 seconds in length, future adaptability to longer content remains seen.

While Google boasts that its watermark is "imperceptible," it is uncertain how resistant it is to removal. Additionally, it is not clear how effectively other platforms can detect and label content with embedded SynthID as AI-generated.

Gemini's capabilities cover all languages and locations where the app is available, making it globally accessible. This move aims to enhance transparency and traceability of AI-generated content across different media.

The integration of coherently tagging AI-generated material is crucial in curbing the spread of deepfakes. Google employs C2PA metadata in its AI systems, promoting standardization in how AI-created content is recognized and tagged online.

 #

--

Yoooo, this is a quick note on a link that made me go, WTF? Find all past links here.