Google Introduces Gemini Omni, a Multimodal AI That Knows the World

Google unveiled Gemini Omni at I/O, a multimodal AI video model that can generate and edit lifelike videos from text, images, and existing video.

The company said it will launch with SynthID watermarking safeguards and show up in Gemini and other Google products, including Flow and YouTube Shorts.

Google plans to open access to developers and enterprise users through APIs later, with Gemini Flash arriving first and Gemini Pro following.

The launch expands AI video creation and editing capabilities, while renewing concerns about realism, manipulation, and content authenticity.