iOS 13 beta includes an ARKit-based FaceTime feature that makes it look as if users are looking into the camera even when looking at the image on the screen
Jon Porter / The Verge :
Context & Ripple Effects
This feature is the consumer-facing payoff of a pipeline Apple has been building since ARKit 2 introduced improved face tracking in iOS 12: the same depth-camera sensing that powers shared AR experiences now corrects gaze geometry on live video calls, making screen-watching read as eye contact.
It also fits the larger arc of iOS 13 itself — the same release cycle whose GM builds later surfaced references to stereo AR headsets and apps. A capability that ships first as a FaceTime nicety is simultaneously proving out the face-tracking layer a face-mounted AR product would need.
First-order effects
- FaceTime users on iOS 13 beta get synthetic eye contact by default in video calls, changing how presence reads on both ends of every call without any user setup.
Second-order effects
- The iPhone's front camera is further established as a general-purpose face-capture instrument — the direction Epic leaned into with its Live Link Face app streaming facial expressions into Unreal Engine — pulling pro animation workflows onto consumer phone hardware.
Third-order effects
- If the pattern holds, Apple ships face-sensing capabilities as communication conveniences while maturing them for a stereo AR headset, so the phone becomes the development and calibration platform for the next device category.
The trend: Smartphone face tracking is migrating from single-feature novelties like gaze correction toward the sensing stack for face-mounted AR.