Tools · The Signal
Franklin AI tests Gemini 3.8 Live’s new real-time avatar layer
Gemini 3.8 Live pushes conversational AI toward persistent on-screen characters that can see, listen, speak and remain visually present while tools run in the background.
Reporting updated Sep 26, 2026
Gemini 3.8 Live pushes conversational AI toward persistent on-screen characters that can see, listen, speak and remain visually present while tools run in the background.
What Happened
Google released Gemini 3.8 Live with Live Avatar on September 24, and Franklin AI followed with a hands-on YouTube test on September 25.
Why It Matters
This moves real-time generative characters closer to usable entertainment interfaces. The same ingredients behind enterprise avatars—visual awareness, fluid dialogue, expressive video and asynchronous tools—could support interactive hosts, character performances and NPC-like experiences.
VEYR Angle
The interesting question is what happens when a live AI character is treated as entertainment rather than a support agent. Franklin AI’s test gives VEYR a creator-side look at a capability Google is positioning primarily for enterprise use.