Google launches Gemini Omni Flash to turn text, images, and audio into video

In short: Google introduced Gemini Omni, starting with Omni Flash, a tool that can make and edit short videos from text, images, audio, and video.

What happened

Google announced Gemini Omni at its Google I/O conference. It is a new family of AI models that can work across different types of media, including text, images, audio, and video.

The first release is Gemini Omni Flash. Google says Flash can create a 10 second video by taking mixed inputs, like a few photos, a short audio clip, and written instructions. Instead of simply gluing pieces together, the system tries to keep the result consistent, like a single scene that makes sense.

Omni can also do simple photo editing through written commands. For example, you can type what you want changed instead of using traditional editing tools.

Google is also adding a way to make videos using your own digital avatar, which is a computer-made version of you (like a reusable puppet of your face and voice). To reduce misuse, users must record themselves and speak a series of numbers during setup. Google says Omni videos will include SynthID, a digital watermark that helps show the video was created with Google’s tools.

Omni Flash is rolling out to the Gemini app, YouTube Shorts, and Google’s creative tool Flow. Google also plans to offer access through an API, which is a way for other apps to connect to it, in the coming weeks.

Why it matters

Tools that create video from everyday inputs could make it easier for regular people, small businesses, and creators to produce short videos quickly. At the same time, features like avatars raise concerns about fake videos, so watermarking and identity checks will matter in how widely people trust what they see.

Source: TechCrunch AI