Google Gemini will be able to control a smartphone; new functionality was discovered in the code

Artificial intelligence from Google will soon be able to control your gadgets

More advanced AI models are gradually moving beyond simple tasks, and in the near future they will start interacting directly with mobile devices. Google is already working on such functionality as part of Project Astra and promises to show early results soon.

What we know about the new feature
- In the beta version of the Google app (version 17.4.66) a mention of an upcoming capability was found on *Android Authority*, currently referred to in code as “bonobo.” The official name will likely be “screen automation” – screen automation.

- The feature will allow simple actions: making purchases in online stores and ordering taxis. A link for details only points to the general Gemini support page, so specifics remain a placeholder.

How it will work
1. Launch

When the feature is first enabled, the user will see a warning advising them to closely monitor the AI agent’s actions. It reminds that “Gemini can make mistakes,” and responsibility for its behavior lies with the user.

2. Security

Google emphasizes that with *Keep Activity* enabled, screenshots will be available to “trained moderators.” Therefore it is recommended not to enter confidential data or payment information into the app. It is still unknown whether Gemini will automatically stop such input or if the user must do it manually.

3. Orders

When purchases are made via AI they will appear in the app under “My orders.” Currently that link opens an empty page – the functionality is not yet ready for public use.

Which apps are already being tested
- Food and goods ordering

- Transport services (taxis)

- Creation of 3‑D avatars

- Adding context from Google Maps

There are no exact lists of applications that Gemini will be able to control yet. It is expected that Google will clarify the list of supported services in the near future.

In short: Google is actively developing a screen automation feature that will allow AI to perform simple actions on your device. Details remain unclear, but it’s already clear what scenarios can be implemented: from online shopping to ordering taxis. Users should closely monitor how Gemini works and follow security recommendations.