What you need to know
- Code files in the Google app hint that the company is working on video generation features for Gemini.
- Video generation in Gemini could leverage an existing Google model, like how Imagen 3 powers image generation in Gemini.
- Currently, Google offers multiple avenues for making AI-generated videos, including the Veo 2 model.
Google offers many artificial intelligence features outside of Gemini, its conversational chatbot that can serve as Android's digital assistant. However, many of them make their way into the Gemini app in some form or another. That could be the case with Google's video generation features, according to an APK teardown from Android Authority.
Version 16.6.23 of the Google app for Android includes multiple references to video generation in relation to Gemini. Typically, Gemini features appear as code in the Google app before making their way to the Gemini app, as the two apps are closely related. Additionally, the codename for Gemini is "robin," and the code file names below directly link "robin" to "video generation" and "videogen."
assistant_robin_video_generation_status_banner.xml
assistant_robin_videogen_status_banner_background.xmlThis is as good of a sign as any that Google is working on video generation features for Gemini. Along with the file names above, Android Authority found the following code strings in relation to the video generation feature for Gemini:
"Working ......"
"We will notify you when it's ready"That doesn't reveal a whole lot about what video generation will look like, or when (or if) it'll arrive publicly. Google already has some video generation infrastructure built as part of its Google AI suite. Veo 2, for instance, is a "state-of-the-art video generation model" available through VideoFX by Google DeepMind. Some of its other video generation tools include Lumiere and VideoPoet.
It's possible that these become integrated with the Gemini app in the future, similarly to how Google integrated Imagen with Gemini for image generation. Last year, Google brought the latest Imagen 3 model to the Gemini app. That means you can ask Gemini to generate an image of a dog and it'll do it all in the app, without needing to use a separate Google AI service. That could be the same experience with video generation if Google eventually integrates it with Gemini.
Google also has Google Vids, a Workspace service that helps users create and edit video. This kind of functionality probably wouldn't make its way to the Gemini app — it's more akin to a Gemini in Workspace feature — though it is possible.
For now, this is mostly speculation. The code snippets and files clearly point to Google experimenting with video generation in Gemini, but the specifics are unclear, and it isn't working at the moment.
