Google adds animated avatar feature to Gemini, enabling speech in 97 languages

Summary

Google introduced the 'Live Avatar' feature in the Gemini 3.8 Live update. This feature offers real-time animated digital characters with lip synchronization and facial expressions. The characters can chat in up to 97 languages while simultaneously displaying information on the screen. The feature is currently available only to Gemini Enterprise customers.

Why it matters

This move takes AI interaction beyond text and voice, transforming it into a multimodal and visually concrete experience. This could fundamentally change corporate customer service, education, and multilingual communication. By enabling language switching during conversation and providing visual responses on-screen, Google is targeting a segment where language barriers and information density are key issues. The launch exclusively for enterprise customers signals a B2B-focused revenue model rather than a consumer release.

Highlights

• Lip synchronization and facial expressions work in 97 languages, including Arabic; users can switch languages mid-conversation.
• The avatar can display visual information on the screen while speaking, combining dialogue with data presentation.
• Outputs include an invisible SynthID watermark and identity protection measures to prevent misuse.

Critical take

The fact that the feature is limited to enterprise customers and that independent benchmarks or latency data have not been shared makes it difficult to assess whether the 'near real-time' claim holds up under real-world network conditions. Additionally, the article lacks details on pricing, hardware requirements, or comparisons with existing video-LLM and avatar platforms. Consequently, the practical value proposition remains largely unverified.