# Look and Talk: Seamless AI Assistant Interaction with Gaze-Triggered Activation

arXiv: https://arxiv.org/abs/2504.09296 · HTML: https://mlsanity.com/paper/2504.09296/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 12 Apr 2025.

## Built on by (most cited first, 1 of 1)

- [Do Audio LLMs Listen Before They Act? Diagnosing Acoustic-Context Gating in Voice Agents](https://mlsanity.com/paper/2609.32536/) (2026)

## Abstract (arXiv)

Engaging with AI assistants to gather essential information in a timely manner is becoming increasingly common. Traditional activation methods, like wake words such as Hey Siri, Ok Google, and Hey Alexa, are constrained by technical challenges such as false activations, recognition errors, and discomfort in public settings. Similarly, activating AI systems via physical buttons imposes strict interactive limitations as it demands particular physical actions, which hinders fluid and spontaneous communication with AI. Our approach employs eye-tracking technology within AR glasses to discern a user's intention to engage with the AI assistant. By sustaining eye contact on a virtual AI avatar for a specific time, users can initiate an interaction silently and without using their hands. Preliminary user feedback suggests that this technique is relatively intuitive, natural, and less obtrusive, highlighting its potential for integrating AI assistants fluidly into everyday interactions.
