Krux
May 21, 2026
Google's Gemini Omni Edits Video Through Conversation
Published: May 21, 2026 at 12:23 AM
Updated: May 21, 2026 at 12:23 AM
100-word summary
Google launched Gemini Omni Flash, a model that generates and edits video through natural language prompts. You can tweak camera angles or swap backgrounds across multiple turns while keeping characters and physics consistent. It works with any input type: feed it images, audio, text, or existing video and it'll blend them into new footage. The model hits YouTube Shorts for free this week, with paid access rolling out today through Gemini's app. All output carries an invisible SynthID watermark to flag AI-generated content. Google clearly sees video creation as the next battleground for consumer AI, not just chatbots answering questions.
What happened
Google launched Gemini Omni Flash, a model that generates and edits video through natural language prompts. You can tweak camera angles or swap backgrounds across multiple turns while keeping characters and physics consistent. It works with any input type: feed it images, audio, text, or existing video and it'll blend them into new footage. The model hits YouTube Shorts for free this week, with paid access rolling out today through Gemini's app. All output carries an invisible SynthID watermark to flag AI-generated content.
Why it matters
Google clearly sees video creation as the next battleground for consumer AI, not just chatbots answering questions.