Google has introduced a new feature called Gemini 3.8 Live, which allows users to interact with an AI model while watching an animated AI persona respond in real time. This feature, called "Live Avatar," includes realistic lip-syncing and facial expressions that change based on the conversation. However, it is currently only available to customers using the Gemini Enterprise version of the AI model. The Live Avatar can switch between 97 different languages without losing video quality or causing visual inconsistencies. A video released by Google shows the avatar speaking in both English and Japanese, with its mouth movements perfectly matching the spoken words in each language. Additionally, the avatar can display relevant information on the screen as it speaks, making the interaction more dynamic and informative. This update comes just a week after Google unveiled the Gemini 3.8 Live model, which is capable of processing visual inputs almost instantly. Google plans to provide a library of pre-designed avatars for users to choose from, while also allowing organizations to create custom avatars tailored to their needs. To ensure responsible use, Google has added an invisible SynthID watermark to the output of Live Avatar. This watermark helps identify that the content was generated by AI, alongside other safeguards aimed at respecting individual identities and preventing misuse.