Gemini Omni is Google’s new AI video-editing model in Flow. Rather than simply placing you over a new background, like a traditional green screen, it can take real footage you upload and intelligently rework it.

You can change the environment, add or remove elements, or even alter aspects of your appearance — while aiming to preserve the movement, character and overall integrity of the original clip. At its best, it feels less like a composite and more like the footage was filmed in that world from the start.

In the second clip in the montage below, you can see the original footage I filmed on my phone.

What I love about some — though not all — of the results is that they go beyond basic green-screen compositing. I’ve been cleanly extracted, curls and all (never easy to key!), and the AI has added lighting — whether sunlight or softer global illumination — that helps blend me naturally into the new scene.

My hope is that the moderation and review process becomes a little less restrictive. Quite a few clips have been rejected during the initial checks, which can make experimenting feel frustrating.

Still, the potential is exciting. Could this breathe new life into old footage? With AI moving as quickly as it is, we may have an answer within a year or two — perhaps even sooner. I can’t wait to see the restorations and reimagined clips people start sharing online.