AI Meeting Summary

This meeting focused primarily on logistics for an upcoming in-person gathering in London, with participants confirming their travel arrangements for Wednesday and Thursday. Fabien Bénétou demonstrated multi-user collaboration examples in XR, showing how users can interact asymmetrically (some in headsets, others on desktop) and how collaboration sessions can be replayed using timestamped messages. The discussion evolved into deeper topics about digital identity, truth in telepresence, and the philosophical question of what constitutes a “real self” in increasingly mediated digital interactions. Brandel Zachernuk shared insights from recent W3C conferences focusing on cybernetic avatars and spatial identities, while the group touched on concerns about AI-generated content displacing human authors and the unexpected ways personal data can resurface through services like ancestry DNA testing.

AI Important Points and Notable Items

Fabien Bénétou emphasized that not everyone wants or can be in XR, highlighting the importance of asymmetric collaboration where some users work in headsets while others contribute from desktop environments

The replay functionality for collaboration sessions was demonstrated, allowing late joiners to understand what happened by replaying timestamped messages rather than seeing an unintelligible mess

Brandel Zachernuk shared his early 2000s startup experience using IRC as a transport layer for co-present retail experiences, recording IRC logs and using bots to replay them for simulating multiple users

Ken Perlin’s lab has been building WebXR interfaces for years where clients can be either desktop or headset, focusing on understanding equivalents between modalities in multi-user environments

Fabien Bénétou proposed integrating Delta Chat or similar open-source interoperable chat systems to allow users to contribute text, code snippets, or documents without needing to load the 3D environment

Discussion of how social media users often present filtered versions of themselves to the point where the filtered version may feel more “real” than their actual appearance

Frode Hegland proposed an artistic concept for a face-changing app that works as a mirror, allowing users to see themselves as different races or genders to build empathy and understanding

Recent W3C Olive Group conference in Kobe, Japan featured discussions on cybernetic avatars, robotic teleoperation, telepresence, and experiencing memories

Apple Vision and similar companies are conservative about transmitting and presenting avatars due to concerns about interception and misrepresentation

The Alliance for Open General Symposium brought together USD experts from Pixar and others to discuss responsible ways of declaring that 3D representations accurately represent actual people or things

Concern raised about a potential future where only wealthy people can afford to read books written by humans, while those with less money must read AI-generated content (from article in Independent newspaper)

A friend discovered she has a brother through Ancestry DNA testing years after the test, highlighting how personal data can resurface unexpectedly

Frode Hegland noted that AI sales calls are now so common and sophisticated that only specific accents help distinguish human callers from AI

Reference to Iain M. Banks’ Culture novels where people in the far future choose slightly quirky appearances rather than conventional beauty, with the observation that digital spaces already mirror this tendency