A recent study shows, that only ๐ณ๐ถ๐๐ฒ ๐บ๐ถ๐ป๐๐๐ฒ๐ ๐ถ๐ป ๐ฉ๐ฅ ๐ฐ๐ฎ๐ป ๐ถ๐ฑ๐ฒ๐ป๐๐ถ๐ณ๐ ๐๐ผ๐๐ฟ๐๐ฒ๐น๐ณ with a high percentage. You donโt have to be logged in, authenticated, or somehow identified. Only your movements, the time pattern of your gestures, or the way you control your avatar is enough (see https://lnkd.in/euRE6xyw).
The authors show, that their system is able to identify an user with 95% accuracy out of a pool of 511 participants when trained less than 5 minutes.
This outlines, that our movements, behavior, actions and interactions in virtual spaces are not anonymous and can be used to inference our real identity. Thatโs kind of โFace-IDโing you Avatarโ without knowing.
This is quite worrying and scary in ad least two perspectives:
1. ๐๐๐ฒ๐ฟ๐ ๐ฎ๐ฟ๐ฒ ๐ป๐ผ๐ ๐ฎ๐๐ฎ๐ฟ๐ฒ ๐ผ๐ณ ๐ฏ๐ฒ๐ถ๐ป๐ด ๐ถ๐ฑ๐ฒ๐ป๐๐ถ๐ณ๐ถ๐ฎ๐ฏ๐น๐ฒ in anonymous worlds
2. identification in ๐ฉ๐ฅ ๐ฟ๐ฒ๐พ๐๐ถ๐ฟ๐ฒ๐ ๐ฐ๐ผ๐ป๐๐ฒ๐
๐๐๐ฎ๐น ๐ถ๐ป๐ณ๐ผ๐ฟ๐บ๐ฎ๐๐ถ๐ผ๐ป and such contexts can easily be ๐บ๐ถ๐๐ถ๐ป๐๐ฒ๐ฟ๐ฝ๐ฟ๐ฒ๐๐ฒ๐ฑ leading to wrong results.
In terms of education and public awareness, this is a hot topic to discuss! Also for vendors or virtual applications, this is an important finding on which various use cases could be built.
Iโll utilize the upcoming events in May, like the MEWS โ The Metaverse Entertainment Worlds or Identiverse, to provide more insight and showcase possible solutions and will keep you updated here.