Gemini Live: Chat Naturally with Google's Voice AI
Following its recent conference, Google is rolling out Gemini Live, the voice mode of its chatbot, allowing you to converse with artificial intelligence as if it were a human being—a sort of Siri boosted with AI!
At the Made by Google conference held on August 13, Google unveiled its new range of Pixel devices, as well as its latest advances in artificial intelligence. After presenting the new features of its mobile AI, Gemini, with varying degrees of success, the Internet giant put Gemini Live in the spotlight. This feature, which promises to let you interact with the personal assistant as if it were a human, was initially introduced last May at Google I/O. It's described by the company as a "mobile conversation experience that allows you to have natural conversations with Gemini," according to a blog post. You will be able to ask it questions on various topics, have spoken conversations, and even interrupt it mid-answer to delve deeper into a particular point—all while benefiting from its advanced voice synthesis and responsiveness.
Gemini Live: A Personal Voice Assistant Enhanced by AI
The presentation of Gemini's new features didn't go smoothly. There was a small hiccup during the demonstration of the chatbot's integration into the Tasks, Calendar, and Keep applications. Then came Gemini Live. To demonstrate its functionality, Jenny Blackburn, Google's vice president of AI, asked: "My niece and nephew are coming over this weekend, and I need some ideas for something fun and educational that we could do together. I was thinking maybe a chemistry experiment, something a little magical." The assistant suggested creating a homemade volcano, a classic chemistry experiment, before asking if she wanted to hear other suggestions.
The Google executive replied affirmatively, and the AI then proposed making a lava lamp with heat-reactive ink. When Jenny Blackburn wondered if the workshop would be too messy, Gemini explained that she would simply need to cover the work surface. Finally, the Google executive asked for name suggestions for the experiment, and the AI quickly suggested "The Spy Training Academy" or "The Secret Message Lab." In short, it's a true conversational assistant! Google emphasized that you can interrupt Gemini Live mid-response to redirect it or change the subject, or even pause the conversation. It's also possible to use the assistant in the background or with your smartphone locked, similar to how you would use a call. In terms of interface, no photos or text appear on the screen to avoid disrupting the conversation.
To offer more customization, Google provides no fewer than ten available voices, each with different characteristics and a fairly natural tone. Three of these voices—Vega, Dipper, and Ursa—were presented during the conference. For now, Gemini Live is only available in English and exclusively for subscribers to Gemini Advanced, which costs $20 per month on Android smartphones. The AI will be available in other languages and on iOS in the coming weeks.
This launch comes just days after the deployment of ChatGPT's advanced voice feature in beta. Reserved for certain ChatGPT Plus subscribers, it offers, according to OpenAI, "more natural and real-time conversations, which can be interrupted at any time." The competition between the two companies continues to intensify as they race to lead in artificial intelligence.