Google is heavily invested in advancing Gemini, and the forthcoming update to its application is set to enable users to accomplish numerous tasks without needing direct manual input on their smartphones. Previously, Gemini’s capabilities were limited to fundamental operations, such as text correction and solving mathematical problems. Now, the artificial intelligence is poised to take control of installed applications.
This feature, internally codenamed “bonobo,” was spotted within the beta version of Google 17.4 and is currently accessible in an experimental capacity through Google Labs. It will empower Gemini to autonomously interact with the applications installed on the device. For instance, a user could arrange an Uber ride or order food via Uber Eats without manually launching the respective apps or setting parameters.
Initially, this automation functionality will only support a select group of applications where users are most likely to employ it frequently. Google is expected to roll out this feature first within its proprietary applications before expanding compatibility to third-party offerings. To utilize this, the device must be running Android 16 QPR3 or a later version.
Users are cautioned that Gemini is fallible and may make errors. Any actions executed by the agent on a user’s behalf will ultimately be the user’s responsibility. Should an unwanted outcome occur, the user retains the ability to halt the ongoing process and resume direct control. Furthermore, Google emphasizes data privacy: users are discouraged from sharing login credentials or payment details with Gemini. If the “Activity Saving” option is enabled, specified experts will have the capacity to examine screenshots of Gemini’s operations for service refinement purposes.
This new capability bears a resemblance to features already present in Google Chrome, where Gemini 3 (on Pro and Ultra versions) facilitates the automation of monotonous activities, such as automatically populating online forms. This level of automation is now being extended to the mobile iteration of Gemini.
Considering these forthcoming enhancements, users will be able to simply delegate commands to Gemini for arranging transport, ordering meals, and completing other repetitive errands. Manual phone interaction for performing these specific actions will become largely superfluous.
About the Author
