Home / Technology / Google's AI Omni Can Now Generate Realistic Videos
Google's AI Omni Can Now Generate Realistic Videos
19 May
Summary
- Gemini Omni is a multi-modal AI capable of understanding diverse inputs.
- It can generate realistic videos with accurate physics and context.
- Users can conversationally edit videos, changing backgrounds or styles.
Google announced Gemini Omni at Google I/O 2026, describing it as a significant step toward artificial general intelligence (AGI). This new AI model is designed to generate "anything" from any output.
The initial release, Gemini Omni Flash, excels in video generation. Unlike previous tools, Omni is multi-modal, processing inputs like text, audio, images, and video to create interactive worlds with real-world knowledge.
Gemini Omni aims to produce more realistic videos through accurate physics and contextual understanding of prompts. It also introduces conversational video editing, enabling users to alter aspects like backgrounds or styles through dialogue.
Features like creating digital likenesses via Avatars are under testing for responsible deployment. All generated videos will include a SynthID watermark to confirm their AI origin.
Gemini Omni Flash is available today for paid Google AI Plus, Pro, and Ultra subscribers via the Gemini app and Google Flow. It will also launch on YouTube Shorts and the YouTube Create app this week.
