Technology briefing

Google launches Gemini 3.8 Live Avatar for Enterprise with lip-synced AI faces

Google said Gemini 3.8 Live’s new Live Avatar feature lets Enterprise customers talk with animated AI personas that lip-sync and emote in real time across 97 languages, with SynthID watermarking and custom avatar options—available to Gemini Enterprise, not broadly to consumers yet.

By US Brief desk · Updated 2026-09-25T05:21:00-07:00

Editor’s note: Product facts follow The Verge (Emma Roth, Sep 24, 2026, 7:59 p.m. UTC) and Google’s Gemini Live Avatar announcement as summarized there. Enterprise-only availability, 97-language claim, lip-sync/emotion behavior, on-screen tool use, SynthID watermarking, and custom avatars are attributed to Google via that reporting—not independent US Brief testing. Distinguish Live Avatar from the separate Gemini 3.8 Live model reveal about a week earlier.

Google on Thursday rolled out Live Avatar for Gemini 3.8 Live, an Enterprise feature that puts a talking, lip-synced animated face on live AI conversations—continuing the company’s push to make agentic assistants feel more present while keeping the first release off the broad consumer Gemini app.

What shipped: According to The Verge’s Sep 24 report, Live Avatar can lip-sync and show facial expressions during conversations, switch among 97 supported languages without (Google says) degrading video fidelity, and pull information on-screen while speaking. Organizations can pick preset avatars or create their own.

Who gets it: Google is offering the capability to Gemini Enterprise customers first. That matters for IT buyers and contact-center experiments, but it means most consumer Gemini users will not see Live Avatar in the same launch window.

Safety claims: Google says Live Avatar output carries an invisible SynthID watermark and includes safeguards meant to “respect identity,” per Verge’s summary of the company announcement. Watermarking aims to help platforms and users spot synthetic media; identity safeguards are intended to reduce unauthorized mimicry—though independent audits of both were not part of launch-day coverage.

Why it matters: Real-time talking avatars sit at the intersection of customer support automation, deepfake risk, and workplace AI adoption. A lip-synced Enterprise agent may cut call-center costs and increase engagement metrics; the same stack raises questions about disclosure, consent, and whether users know they are speaking to a synthetic persona.

Context in Google’s week: Live Avatar lands about a week after Google’s Gemini 3.8 Live model messaging around near-real-time visual inputs, and a day when Google also drew attention for Project Suncatcher orbital TPU plans—together painting an infrastructure-plus-interface AI offensive.

Uncertainty and competing views: Launch blogs emphasize fluency and safety; skeptical coverage (Engadget, TechRadar) highlights the “creepy” uncanny-valley risk and limited consumer access. Performance across languages, latency on real Enterprise workloads, and how often SynthID is actually detected by third parties remain unproven in public tests.

What to watch: Enterprise customer case studies; any consumer Gemini timeline; regulators’ interest in synthetic-media labeling; and whether rivals (Meta’s Muse realtime avatar research demos, OpenAI voice mode) answer with similar face-forward agents.

Follow US Brief: Bluesky · X · RSS

Full interactive version requires JavaScript. https://us-brief.com/article/tech-gemini-avatar

National & world Daily briefing

Subject briefs,
national scope.

Pick a subject and get a lean, dated briefing—US News, Politics, Technology, World, or Finance. AI-assisted digests from the US Brief desk with sources.

Back to Technology briefing Technology

Google launches Gemini 3.8 Live Avatar for Enterprise with lip-synced AI faces

Google said Gemini 3.8 Live’s new Live Avatar feature lets Enterprise customers talk with animated AI personas that lip-sync and emote in real time across 97 languages, with SynthID watermarking and custom avatar options—available to Gemini Enterprise, not broadly to consumers yet.

By US Brief desk · Updated 2026-09-25T05:21:00-07:00

AI-assisted digests from the US Brief desk. Synthesizes reputable reporting with cited sources; not a wire service. Corrections welcomed.

Google on Thursday rolled out Live Avatar for Gemini 3.8 Live, an Enterprise feature that puts a talking, lip-synced animated face on live AI conversations—continuing the company’s push to make agentic assistants feel more present while keeping the first release off the broad consumer Gemini app.

What shipped: According to The Verge’s Sep 24 report, Live Avatar can lip-sync and show facial expressions during conversations, switch among 97 supported languages without (Google says) degrading video fidelity, and pull information on-screen while speaking. Organizations can pick preset avatars or create their own.

Like US Brief? You can support it with a tip.