Google Search is shifting from typed queries to real-time voice conversations. With Gemini 3.8 Live, users can now tap the new "Live" icon in the Google app, ask questions out loud, and get spoken answers from the AI. The screen shows clickable web links alongside the audio, so users can dig deeper if they want.
Gemini 3.8 Live and its Extended Thinking variant can automatically switch across 97 supported languages during a single conversation, enabling seamless multilingual interactions.
Rajan Patel, VP of Engineering for Search at Google, confirmed the launch: “New Gemini audio models just dropped - 3.8 Live is now powering real-time conversations in Search Live.” The rollout is available to all users in Search Live, with developer access through the Gemini API and Google AI Studio, and enterprise access in private preview via Gemini Enterprise. Google says the model supports more natural dialogue and fluid language switching mid-session.
For users who want a written record, a "transcript" button shows the full conversation in text. You can review or continue the exchange by typing, and every interaction is saved in AI Mode history for easy reference. This mix of voice and text is meant to keep users engaged and make it easier to revisit or build on past searches. Gemini 3.8 Live is built as a native speech-to-speech model, supporting real-time voice conversations and visual input, as described in Google’s developer documentation.
Every audio output generated by Gemini 3.8 Live is embedded with a SynthID watermark, a signal designed to make AI-generated audio detectable and traceable for authenticity.
For publishers and marketers, technical reliability is a concern. Recent issues, like the Google Analytics data loss, show how quickly platform changes can affect the digital landscape. With Gemini 3.8 Live, any model update or glitch could instantly impact millions of search sessions and the visibility of publisher content.
Content creators and SEO professionals now face new challenges. Search Live’s AI-generated answers include web links in the audio interface, but Google hasn’t explained how it chooses or ranks those links. The ability to switch languages mid-query could also change traffic patterns, making it harder to predict or optimize for user intent. As Gemini evolves, publishers will need to watch not just rankings, but also how their content appears and is delivered in Google’s AI-powered environment.
Google’s rapid rollout of Gemini 3.8 Live shows its intent to lead the next phase of search. By focusing on real-time, voice-driven interaction and multilingual support, Google is betting that users will prefer a more conversational, less text-heavy search experience. For publishers, this means adapting quickly to new formats and staying alert as AI-driven changes reshape how audiences find and engage with content.
Google, founded in 1998, now handles over 8.5 billion searches per day worldwide. Its Gemini AI models, introduced in 2023, are now central to search, productivity tools, and cloud services. As of 2026, Google’s parent company Alphabet reported annual revenues over $300 billion, with AI-driven products making up a growing share of user engagement and revenue.
