Google Launches Gemini 3.8 Live Avatar for Enterprise, Supporting 97 Languages
Updated
Updated · The Verge · Sep 24
Google Launches Gemini 3.8 Live Avatar for Enterprise, Supporting 97 Languages
3 articles · Updated · The Verge · Sep 24
Summary
Gemini 3.8 Live now lets Gemini Enterprise users talk to an animated AI persona that lip-syncs and changes facial expressions in real time.
Google said the Live Avatar can switch across 97 languages without visual drift or lower video fidelity, and can display information on-screen while speaking.
Enterprise customers can choose from preset avatars or build custom ones, while Google adds invisible SynthID watermarking and identity safeguards to the output.
The release follows Google’s unveiling about a week ago of Gemini 3.8 Live, a model the company says can process visual inputs in near real time.
Could the ability of these enterprise avatars to process live camera feeds inadvertently expose sensitive company data to unforeseen surveillance risks?
With deepfake scams costing millions, will Google’s invisible watermarks truly stop its ultra-realistic avatars from becoming the ultimate corporate phishing tool?
If an AI avatar flawlessly mimics human emotion while running background tasks, how will this redefine our baseline for workplace trust?