Google adds a talking animated face to Gemini 3.8 Live for businesses

In short: Google has added a “Live Avatar” option to Gemini 3.8 Live that lets people talk with its AI while watching an animated face respond in real time.

What happened

Google announced an update to Gemini 3.8 Live that adds an on-screen animated character called Live Avatar. Instead of only reading text or hearing a voice, users can watch a face talk back while they chat.

Google says the avatar can lip-sync, meaning its mouth movements match the words being spoken, and it can show different facial expressions during a conversation. Google also says the avatar can switch between the 97 languages Gemini supports without the video getting worse or the face slowly drifting out of alignment.

In a video shared by Google, the avatar speaks in both English and Japanese, and the mouth movements line up in each language. Google says the avatar can also show information on-screen while it talks, like pulling up a helpful note during the conversation.

For now, Live Avatar is only available to Gemini Enterprise customers, which are typically businesses and larger organizations. Google says it will offer a set of preset avatars to choose from, and it will also let organizations create their own avatars.

Google adds that Live Avatar output includes an invisible SynthID watermark, which is a hidden marker meant to help identify AI-made content. The company also says it is adding safeguards designed to “respect identity.”

Why it matters

Giving an AI a face can make it feel more like talking to a person on a video call, which could be useful for customer support, training, and internal help desks. It also raises practical questions about trust, since a realistic talking face can make AI content easier to mistake for a real person.

Source: The Verge AI