Embodied gestures improve ai latency feedback in virtual reality
Wait for It: A Component-Level Analysis of Latency Feedback Mechanisms for LLM Agents in VR
Human-Computer Interaction
Summary
AI responses can be slow in virtual reality, which makes the experience less fun. The authors tested three ways to show users the AI is thinking: saying small words like 'um...', showing loading spinners, or using body movements that look like thinking. They found that body movements helped users feel like the AI was more alive and social. Loading spinners made people feel less immersed because they reminded them of computers, not humans.
What this means in practice
- •For virtual reality developers: Design AI agents with natural gestures to improve user engagement and maintain immersion despite response delays.
- •For game designers: Use embodied latency feedback in interactive characters to enhance social realism and player experience during AI-driven delays.
Authors
Pratham Arora, Abhinav Lodha, Siddharth Siddharth
Abstract
Delays in AI responses pose a significant challenge to immersive Virtual Reality experiences. While previous research demonstrates that feedback can reduce the perception of waiting, we lack understanding of which types work best and why. We conducted a within-subjects study with 18 participants to systematically compare three isolated feedback approaches during VR agent interactions: verbal fillers (e.g., um..., hmm...), visual indicators (loading spinners), and embodied gestures (thinking animations). Our results revealed that embodied gestures significantly outperformed visual loading indicators across multiple dimensions, making agents feel more lifelike and socially present. Interestingly, explicit visual feedback like loading spinners actually degraded perceived humanness and broke immersion by forcing users to shift from social to technical mental models. These findings suggest that natural, behavior-based feedback maintains the illusion of conversing with a thinking entity, while technical indicators expose the computational substrate. We present empirically-grounded design guidelines for creating more engaging AI-powered VR agents.