Google’s Gemini Live AI Assistant Gains Screen Highlighting and Deeper App Integration

• Visual Object Highlighting

  Starting August 28, 2025, Gemini Live will allow users to point their camera at objects, and the assistant will visually highlight the correct choice. For example, when faced with multiple tools, Gemini will mark the right one directly on the screen. It can also help with everyday choices, such as selecting between two pairs of sneakers to match an outfit.

  Google confirmed the feature uses a white-bordered rectangle to identify items on-screen, making it easier to distinguish objects in cluttered or visually complex settings. Initially teased at Google I/O 2024, the feature is launching with the Pixel 10 series before rolling out to other Android phones the same week, and iOS in the following weeks.

• Integration into Native and Productivity Apps

  Gemini Live now integrates with Google Calendar, Keep, and Tasks, available immediately, letting users create shopping lists, manage appointments, or set reminders. With the latest expansion, Gemini is also being embedded into Phone, Messages, and Clock, enabling hands-free calls, texts, and alarms. Google Maps support is being deepened as well, allowing seamless combinations of navigation and messaging.

  This expanded footprint transforms Gemini Live from an app-level assistant into a core, system-wide companion across communication, scheduling, and mobility.

• Seamless Flow of Actions

  Gemini Live supports fluid task-switching in natural dialogue. For instance, while checking subway directions in Google Maps, a user can say, “Send Alex a note that I’ll be 10 minutes late,” and Gemini drafts the message instantly without leaving the navigation screen. Similarly, while brainstorming birthday gift ideas, Gemini can seamlessly move from discussion into placing a call or adding reminders.

• Human-Like Audio Model

  Google is introducing a new speech model with significantly improved intonation, rhythm, and pitch, making Gemini sound more dynamic and engaging. The Verge notes that these refinements make interactions feel noticeably less robotic, with Gemini adjusting tone based on emotional context — calmer during stressful conversations, more upbeat during casual ones.

• Customizable Speaking Speed and Styles

  Users gain control over Gemini’s speaking speed, slowing it down for note-taking or accelerating when pressed for time. Gemini can also adopt character voices and accents for narrative tasks, such as recounting history from Julius Caesar’s perspective. These stylistic flourishes highlight Google’s intent to blend utility with creativity.

• Expressive Storytelling Capabilities

  Gemini Live goes beyond factual responses, adding theatrical qualities to narration. With rich accents, pacing, and role-play delivery, the assistant becomes useful in educational settings (e.g., history lessons), entertainment (children’s stories), and creative writing.

• Strategic Rollout and Next Steps

  While some voice and app integrations will phase in gradually, the Pixel 10 series will be the first device family to showcase Gemini Live’s full feature set. Google’s roadmap makes clear its ambition: to evolve Gemini from a Q&A bot into a universal AI assistant that handles communication, navigation, productivity, and entertainment in one conversational flow.

Why This Matters

These upgrades mark a pivotal moment in the race for AI assistant supremacy, directly challenging Apple’s Siri evolution and OpenAI’s conversational platforms. By merging visual awareness, natural voice, and deep app integration, Gemini Live shifts from being a chatbot to a true digital companion. The inclusion of core phone functions (calls, messages, alarms) and visual learning tools extends its role beyond productivity into daily life management. With rollout beginning on Pixel 10 and scaling to Android and iOS, Gemini Live is poised to become one of the most widely adopted, indispensable AI assistants in the mobile ecosystem.

This article was drafted with the assistance of generative AI. All facts and details were reviewed and confirmed by an editor prior to publication.