Filling Silent Streams: How AI Avatars Keep Engagement Alive Without Viewer Comments
📝 Originally published (in Japanese) at forge.workstyle.tech . The Challenge of "Silence" in Unmanned AI Avatar Live Streams When creating a live stream where an AI avatar operates autonomously, the first major hurdle you encounter is the issue of "silence." It’s not that there are no viewers—quite the opposite. Yet the avatar falls silent for long stretches, or ignores comments for tens of…
When creating live streams with AI avatars, silence is a significant challenge. Unlike human streamers who naturally create conversational space, AI fails to do so. To address this, two main issues were tackled: filling the silence when no comments are present and handling response delays when comments do arrive.
The initial approach was straightforward—respond when a comment arrives. However, this didn't work for streams without comments, as the avatar would remain frozen for long periods. This was a design flaw, as live streamers need to keep talking even when no one is commenting.
The solution was to generate speech autonomously when there's no input. The silence threshold of 75 seconds was determined by observing real streams. If no activity occurs for 75 seconds, the avatar generates a topic based on the stream's theme. This threshold considers viewer delay, which is 15-30 seconds, to ensure conversation flow isn't interrupted.
Spontaneous topics are given low priority in the utterance queue, ensuring comments take precedence. Topics are generated from the stream's theme and personality, avoiding repetition by referencing conversation history. This system was tested over 2-hour streams to confirm its effectiveness.
Despite solving the silence issue, the focus shifted to another problem: the 45-second response time to viewer comments. The dominant factor here was platform-level streaming delay, which cannot be reduced despite optimization attempts. Thus, focusing on reducing other factors like chat retrieval and response queue waiting time became necessary for smoother interactions.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
