Google has introduced Search Live, a new voice-based feature for users enrolled in the AI Mode experiment, allowing real-time spoken conversations with its search engine on Android and iOS.
This new capability brings natural language voice interaction to Google's search experience, supported by a customized Gemini AI model.
Liza Ma, Director of Product Management for Search, outlined in an announcement how the feature enables a dynamic, back-and-forth dialogue with Google's AI. Users can tap a new “Live” icon in the Google app to begin speaking to the assistant, which responds with AI-generated audio and offers clickable links for deeper exploration. Conversations can continue fluidly even as users switch between apps, with support for both voice and typed input.
Search Live is part of Google's ongoing AI Mode initiative, a Labs program that gives early access to experimental search features. Unlike traditional voice assistants, which often rely on static or templated replies, Search Live is designed for extended, context-aware interactions. It draws on Google's existing search infrastructure and incorporates advanced voice synthesis and comprehension, creating an experience closer to natural human dialogue.
Technically, the system operates on a customized version of Gemini, Google's large multimodal model architecture, optimized for conversational and search reliability. The model is augmented with Google's query fan-out system, a backend method that broadens the range of sources scanned during a query. This ensures that responses pull from a diverse set of websites and web content, rather than relying solely on AI summaries or single-source data. The result is a hybrid experience combining real-time AI dialogue with direct web exploration.
Users can also access conversation transcripts for reference and continue sessions from their AI Mode history. This continuity feature hints at Google’s broader goal: making search feel less like a series of discrete queries and more like an ongoing personal assistant session, accessible across apps and devices.
The launch comes amid increasing integration of generative AI into consumer-facing products, as tech giants race to redefine search around conversational experiences. For Google, which still dominates the global search market, AI Mode serves as a testbed for ambitious new directions in human-computer interaction.
Google also informed of plans to extend the modality of Search Live beyond voice. A forthcoming update will allow users to show what they’re talking about by using their smartphone cameras in real time, potentially enabling visual-assisted search and richer multimodal queries.
Search Live is only available to those enrolled in the AI Mode Labs program, so interested users should opt in through the Google app. To ensure a secure and private experience, review the Labs settings and manage voice history and data sharing under your account’s privacy controls. While helpful, AI-generated results should still be cross-referenced for accuracy, especially when relying on the information for travel, health, or critical decisions.
Leave a Reply