Google usually makes Live Avatar available in Gemini Enterprise: a visual interlocutor who listens, sees and responds in real time, watermarked by SynthID.
On September 24, Google opened general access to Live Avatar for Gemini 3.8 Live in Gemini Enterprise. The principle: to add an animated visual presence to the speech model, with lip-syncing, facial expressions and management of turns. The system processes audio and visual input simultaneously, executes tool calls asynchronously and supports 97 languages.
In terms of supervision, Google distinguishes between two regimes: a library of pre-designed avatars accessible to all customers, and the creation of personalized avatars, reserved for an authorization list with a verification procedure. All the audio and video streams generated bear an imperceptible SynthID watermark, intended to keep the synthetic content identifiable.
The use cases targeted are customer relations and interactive guided paths. The service is offered with US and European access points, provisioned throughput and corporate compliance commitments – a fairly direct response to the transparency mandates that came into force in Europe at the beginning of August.
Read the full article on the original source
Go to the source →