Google has notified some Android users that its AI assistant, Gemini, will soon gain access to key phone applications, including Phone, Messages, WhatsApp, and Utilities, regardless of whether users have turned off Gemini’s App Activity setting. The update is scheduled to take effect on July 7 and has sparked concerns about data privacy and the transparency of user consent.
New Access Extends to Core Communication Apps
The notification, first highlighted by a Reddit user and later reported by Android Police, was sent via email to certain Android users. According to the message, Gemini will be able to assist with several core phone functions even if a user has opted out of tracking through the Gemini Apps Activity setting. This includes the ability to interact with apps that often contain sensitive information, such as personal messages, call logs, and daily utility usage.
Google stated in the email that users who do not wish to enable these features can turn them off through the app settings page. However, the email did not include specific instructions on where to find that setting or what functionality would be affected by disabling it. The vague language has left some users unclear about the full scope of Gemini’s default behavior and the extent to which app access will continue in the background.
The company has previously explained that when App Activity is turned on, Gemini’s interactions, including prompts and responses, may be stored for up to 72 hours, and some of this data could be reviewed by human evaluators. It is not yet clear whether similar data handling will apply to users who have App Activity turned off, now that Gemini’s access is expanding. Google has not yet responded to media inquiries seeking further clarification on the matter.
AI Integration Renews Privacy Debates
The update has prompted renewed discussion about the privacy implications of AI integration into personal devices. Gemini, part of Google’s push toward what it calls “agentic AI,” refers to AI that can independently perform tasks like scheduling appointments, ordering items, or responding to messages. This functionality requires deeper access to user data to work effectively. That level of access raises questions about what data is collected, how it is used, and whether users have sufficient control over their personal information.
Unlike earlier forms of AI assistants, such as voice-based tools that were typically limited to predefined queries, Gemini’s design supports broader integration. This includes interpreting and acting on content from messaging apps, which are among the most sensitive sources of data on any mobile device. Critics argue that default access to these apps, especially when users believe they have opted out of such features, may erode trust in AI systems and in Google’s privacy policies.
The situation recalls earlier debates over voice assistants and smart home devices, where questions about passive listening and user consent took years to resolve. As AI becomes increasingly embedded in operating systems and app workflows, companies like Google are being called upon to provide clearer options for users who do not want their data used in this way. The lack of detailed opt-out instructions in Google’s notice has only intensified concern.
Questions Remain Ahead of July 7 Rollout
With the July 7 deadline approaching, users are still waiting for clarity on how to manage or restrict Gemini’s access to their apps. Although Google has provided a general opt-out mechanism, its omission of specific steps and definitions has left room for confusion. Until the rollout occurs, it remains unclear how the new permissions will be implemented or enforced on individual devices.
Some privacy advocates argue that this type of system should be opt-in rather than opt-out, particularly for features that interact with personal communications. They point out that many users may not realize Gemini is analyzing content from their messaging apps, especially if no clear prompt or permission request is presented when the feature is activated. These concerns highlight the broader challenge of ensuring informed consent in AI-powered services.
As generative AI continues to expand into everyday mobile functions, the balance between convenience and privacy will be increasingly important. Users are being asked to trust systems that require access to sensitive data to perform tasks efficiently, but that trust must be supported by transparency and control. For now, Google’s handling of the Gemini update suggests that the conversation about AI privacy is far from over.