Skip to content
TopicTracker
From HackerNewsView original
TranslationTranslation

Interactive Avatars

Synthesia's Interactive Avatars are AI-generated video avatars that can speak, gesture, and respond in real time. Designed for enterprise use, they enable interactive video experiences without the need for cameras or actors, supporting applications like virtual presenters and customer support.

Background

- Synthesia is a London-based AI startup (valued at ~$1B as of 2023) that lets users create professional-looking videos with AI-generated human avatars — no camera or studio needed. - An "interactive avatar" is a Synthesia feature that goes beyond pre-recorded video: the AI avatar can respond in real time to what a user says or types, making it suitable for things like customer support, virtual sales reps, or training Q&A. - This builds on Synthesia's existing technology (text-to-video avatars) but adds conversational ability, likely powered by integration with large language models (LLMs) like GPT-4. - The broader context: generative AI has rapidly advanced from creating static images and text to real-time interactive avatars, raising both business possibilities (cheaper, scalable "talent") and ethical questions about deepfakes and digital impersonation. - Synthesia's avatars are created from real actors (with consent) and include safeguards like a ban on political content and mandatory disclosure that the avatar is AI-generated.

Related stories