Google AR&VR collaborated with the University of Central Florida to launch a virtual character library called "VALID", which contains a full set of 210 virtual character models covering seven different races, aiming to promote diversity and inclusion in virtual environments. The move is intended to push virtual reality technology to better reflect real-world demographics and promote fairer, more representative virtual experiences. However, the findings also point to differences in perceptions of avatars of different races, which raises discussions about potential biases and cultural differences in avatar design.
The article focuses on:
Google AR&VR collaborated with the University of Central Florida to release the "VALID" virtual character library, which contains a full set of 210 virtual characters representing seven different races and aims to promote diversity and inclusion. The research results show that Asian, black, and white avatars obtain consistent cognition, while cognitive ambiguity exists for other races. Co-race bias affects avatar perception, and research highlights the importance of participant race for accurate representation.
The results of this study highlight the importance of considering cultural background and potential bias when creating virtual characters, and also provide valuable experience and reference for the development of future virtual character libraries. How to eliminate algorithmic bias and build a more objective and inclusive virtual world still requires continuous exploration and improvement.