IEEE Technology and Society Magazine - June 2018 - 53

Here we distinguish several types of privacy (derived
from [2]), including:
■ Privacy of personal info: Any information that
reveals something about physical, medical, physiological, economic, cultural, or social status.
■ Privacy of behavior: Any information about habits,
activities, choices, etc.
■ Privacy of communications: Any data and metadata
relating to personal communications.
Note that sometimes we accept a loss of privacy in
exchange for security (in the case of security surveillance) or in exchange for useful customization (e.g.,
personalized advertisements). We also, sometimes unwittingly, freely offer up much of our personal information. For example, our mobile GPS location and device
characteristics may be shared ubiquitously, and our
social media posts may have a surprising reach (e.g.,
150 000+ "friends of friends" [3]). Nowadays, as virtual
reality (VR) applications increase in popularity and fidelity
they also threaten to erode our privacy in new ways ranging from knowing how we physically move around to the
patterns of our neural activities [4].
In this article we focus on technology underpinnings
that will help VR participants increase the degree of privacy while immersed in social VR, and builds on our
past research in privacy and gaming analytics [14],
[15]. Though coined quite some time ago, we use the
term metaverse with the same semantics as in Wikipedia [16]: "a collective virtual shared space, created by
the convergence of virtually enhanced physical reality
and physically persistent virtual space, including the
sum of all virtual worlds, augmented reality, and the
Internet." We use the term social metaverse to describe
the above sorts of virtual realities in which a central
purpose is socialization and interaction with other avatars - including both players and non-player characters
(NPC's). Examples of software systems considered
social metaverses today include: Facebook Spaces,
AltspaceVR, Sansar, High Fidelity, and many more.
While the social metaverse may or may not include
capabilities such as gamification, realistic physics, realistic 3D models, user-created content, or in-game economies, it is the complexity and nuance created by the
presence of other avatars (human or not) that most
motivates our work.
Let us define the term "avatar" (or "agent") as a visible
character within the social metaverse, constrained by the
rules of the metaverse. We'll also use the term "user" (or
"player") to connote a human who operates one or more
avatars. Notably, the social metaverse:
■ Is implemented by an engine that provides the computational basis ("rules of the game") for all aspects
of the world including physics, appearance, communication, synchronization, etc. The engine is in
JUNE 2018

∕

sole control of the consistency and durability of
the metaverse.
■ Hosts avatars who cannot hide from the engine
itself (if the engine attempts to surveil or analyze
avatar activities, it may do so) nor can they perform
actions not offered via metaverse API's.
■ Is sometimes editable in the sense that avatars
can affect the virtual world (e.g., create or des troy objects).
What is also true about the social metaverse is that,
just like in the real world, those avatars who most skillfully use the capabilities of the world in the best way
possible may experience a competitive or social advantage over others. We do not consider this to be a nefarious "gaming of the system" but simply using it better.
Avatars, for example, may leverage a metaverse application program interface (API) and perform their own
sort of surveillance and there is no guarantee that their
actions or intents are ethically sound. For example, inmetaverse stalking is a dubious - but often allowable - kind of interaction with these worlds.1 The
metaverse will surely be underpinned by data analytics
(DA) software components and combined with big data
analytics and machine learning in order to provide the
developer with insights into how users employ their
services [4].
The stage, in our opinion, is therefore set for a battle
for privacy within the social metaverse. While it may
seem at present that little is at stake, one should note
that it is possible that a good deal of our future lives
may play out within these metaverses, including performing productive, meaningful work, exchanging
important ideas, and using valuable digital currencies.

Motivation and Current Landscape
We have described how the stage is presently set for a
privacy battle. This section provides more detail and
some examples to corroborate this view. We also survey
some of the academic work in this realm.
In the virtual reality metaverses seen in Hollywood
movies - e.g., "The Matrix" and "Ready Player One" -
participants often experience a level of fidelity indistinguishable from the real world [5]. The antagonists in
these movies (but sometimes also the protagonists)
often have special powers gained from their uniqueness
or some sneaky shortcut. While the hacking of metaverse software underpinnings is an interesting field of
1

There are increasing examples of inappropriate user behavior negatively
affecting service offerings. Toxic players may insult others or threaten to
"throw" the game or deliberately ruin other gamers' experiences in creative ways. In one example, toxic player behavior in the game Overwatch
has delayed new levels and triggered serious re-thinking of processes for
user management. (See: [17].) Other major entities in this space are both
recognizing and starting to address toxicity-related issues. These include:
Facebook [18], Steam [19], Google [20], and Rockstar Games [21].

IEEE Technology and Society Magazine

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - June 2018