Real‑time Avatars Don’t Create Human Connection

Google’s Gemini 3.8 Live uses a raw‑audio pipeline to produce near‑instant, lip‑synced avatars for enterprise customer interactions, but rapid responsiveness and fluent multimodal output still don’t recreate human connection — users report speed and multilingual skill but also persistent ‘no one’s there’ feeling. The technology thus creates smoother, more persuasive interfaces without solving empathy, which matters for trust and disclosure policies. — If fluency substitutes for presence, platforms will need disclosure, consent, and liability rules because people will treat fast, face‑to‑face synthetic interlocutors as social actors even when no human is present.