- By Prateek Levi
- Wed, 04 Feb 2026 11:34 PM (IST)
- Source:JND
Google GeminiWith each passing day companies are taking steps to make AI integration with devices much more seamless and smoother than ever before, and to deliver on the same, Google might be making some new moves. According to reports, Google is inching closer to giving Gemini, its AI suite, the ability to handle tasks directly on Android devices, as per the latest details found in a beta version of the Google app. The new feature being dubbed as "Screen Optimization" would give Gemini the access to some apps on the user's behalf and carry out tasks like booking rides and placing rudimentary orders.
This new capability appears in the 17.4 beta of the Google app, as reported by 9to5Google. The information came to the fore after an APK teardown took place, which essentially examines the app code to get information that would identify what features might show up next in future updates. Although this finding does not confirm anything, it does hint at the direction the tech giant might be thinking of taking next.
How Will Screen Automation Come Into Play?
If and when it does, the beta consists of references to a feature called "Get tasks done with Gemini," which the tech giant has internally named "bonobo." Text strings linked with the feature suggest that Gemini might be able to take hold of certain tasks using screen automation with the apps that support it on the user's device. Basically it would act like an on-screen agent; rather than just following the usual respond-to-questions format, this time the assistant would follow steps that the app stipulates, such as selection options, action confirmation, or even filling out and completing a form. For now the feature is limited to certain apps, which means if the feature becomes a reality, it would not support all apps at once at launch, at least.
Google has also attached a warning, expressing that Gemini can make mistakes, and places the responsibility on the user itself for actions taken. Users can opt out of it and stop the process anytime they want and take things back into their hands for manual processing.
What About Privacy?
Privacy notices in the beta version give a clearer idea of how the system might actually function. When Gemini interacts with other apps, screenshots from those interactions could be reviewed by trained human reviewers, but only if the user has activity tracking turned on. Google says this is done to improve how the service works.
The company also advises users to avoid sharing sensitive information such as login credentials or payment details in Gemini chats. It further recommends not relying on screen automation for emergency situations or for tasks that involve highly sensitive data.
What Else Is There in Beta?
A few more code names have appeared in the beta, being dubbed "Wasabi," which is linked with something called "Likeness." This seems like something that would be handling 3D avatars, which Android XR already employs in Google Meet calls.
These new findings point towards a new direction that the tech giant might take as AI integration with handhelds is catching fire right now. With more assimilation, it could change the ecosystem of how people interact with their devices, but a few concerns like privacy and security will still prevail if the feature ever comes to the fore.