Gemini Omni Flash: Google's AI Video Tool & Watermarks

Google has launched Gemini Omni Flash, a new AI video model now rolling out in the Gemini app, Google Flow, and YouTube Shorts. This tool promises to create short, editable video clips from various inputs like text, images, audio, or existing video. Here's the thing: Gemini Omni Flash is the first release in Google's new "create anything from any input" model family. It generates roughly 10-second videos with synchronized audio. Users can refine these videos through natural, multi-turn chat. What's interesting is that the model accepts mixed inputs, including a script or a reference image, to produce a single coherent video. Early demonstrations showed the ability to request scene changes, style tweaks, and object swaps while maintaining visual continuity. The bottom line: Google states that every Gemini Omni Flash video carries an imperceptible SynthID watermark for verification. This technology aims to make video creation more accessible, but also raises questions about synthetic media abuse and the need for robust content policies.