From Eye Movements to User Context: Building More Personal AI Glasses with Inseye Tiny

AI glasses are becoming multimodal systems. What is still missing is a low-power signal of the user’s state.

Inseye Tiny is designed to provide that signal through eye-movement sensing. By detecting activity related patterns, Tiny can help multimodal AI glasses become more personal and less interruptive.

This talk will focus on the system opportunity: how eye-movement signals can shape both AI behavior and HUD behavior. Instead of always responding the same way, glasses can learn when to wait, when to nudge, when to unfold a notification, when to trigger OCR, when to scroll, and when to ask for confirmation.

Tiny is designed for real eyewear constraints, enabling user context and interface control without full camera-level gaze tracking.