Google is laying the groundwork for more visible controls over AI agents in Android. Android 17 QPR2 Beta 4 includes a new Agents section that is designed to list services able to access information or carry out actions within apps on a user’s behalf.
The page is located at Settings > Security & privacy > Privacy controls. It does not currently show any active agents, but its description indicates that Android will eventually surface the relevant services there, giving users a single place to review agentic activity and permissions.
Early evidence suggests the dashboard will not be limited to Google-built agents or Gemini Spark. It could also collect agents supplied through third-party apps and services, potentially making Android’s privacy settings a system-level oversight point for AI tools installed on a phone.
The work appears connected to Android Halo, a visual indicator Google presented at I/O 2026. Android Halo is intended to signal when an AI agent is operating in the background, similar to the existing indicators for camera and microphone access. Google has said the feature should reach devices later this year.
Google is also preparing an expanded Privacy Dashboard with a focus on AI activity. The planned interface is expected to show what agents are doing, which information they use, and when data leaves the Private Compute Core. Both the Agents page and the expanded dashboard remain in development, with their final behavior and rollout timing yet to be confirmed.
