IEEE Technology and Society Magazine - September 2017 - 77

In capturing human social value, i.e., notions of how
and why things matter to humans, what is needed is a
rich set of theories for how to actualize and operationalize diverse forms of value in computational systems.
Such theories will be essential to realize the vision of
socially-sensitive systems and their design. Our aim is
to produce computational formalizations of a variety of
forms of social value, enabling them to be explicitly represented within the technical side of socio-technical systems, and communicated by them. In doing so, we will
also provide some of the tools required to ensure that
such systems also uphold these social values.

Challenge 3: Formalizing Richer Social Relations
As we have noted, a tremendous leap in the impacts of
the advantages of multiple agents occurs when we add
scope and perspective as well as purposeful and intelligent support by others. These social processes in living
systems come in a wide variety of types and levels of
interaction and coordination, including ones that make
several agents aware of each other and make them perceptive and anticipatory of each other's needs, allowing
them to be able to model and reason about each other's states and goals. These processes allow agents to
share goals and to communicate broadly about the
environment and the system's own states, and to be collaborative and supportive of each other.
Of course there can be many levels of socially sensitive
processes. A guard dog can alert us and guide us; a mentor can tutor the apprentice. Both physically shape the
learner's behavior, pointing out how what they do impacts
their goals. They point out what to look at in the environment, intellectually and emotionally, helping one reason
through a proof or helping one to not overreact to a social
situation, etc. There are many examples of learning, training, mimicry, and teaching, both formal and informal.
As discussed above, we have the beginnings of
social concepts, including relations, already in use or
in research (e.g., trusted for a specific shared goal within a specific context). However, these are typically not
noted as such, explicitly. To develop much more powerful social relationship concepts, we will need to take
on some challenging topics, such as how to build computational systems that exhibit "intentionality" over
relationships, that will help or engage socially, and that
have the relative freedom or choice to engage. How do
we develop the means for communication and interaction among diverse styles of components and processes? How do we generalize the benefits from early forms
of social processes to more advanced ones? For example, how do we go beyond agents sharing a goal or
trusting one another for a given task to the presence of
trust within a culture, a social relationship, or a tribe,
institution or group of individuals?
SEPTEMBER 2017

∕

In answering these questions, the potential exists
to draw on theories of social relations and repeated
interaction in sociology, where the impact of different
qualities of relationship underlying interactions has
been studied extensively. The challenge is therefore to
formalize and operationalize different forms of social
relation. In doing so, it will be important to characterize
different types of social relation in terms of the (potential) benefits they bring, as well as the costs in maintaining
them, which may otherwise be seen as "inefficiencies"
from a purely engineering perspective.

Challenge 4: Harnessing Self-Organization
and Decentralization
With any system of sufficient complexity, it is important
to consider the internal organization of the system,
including how the comprising parts are arranged. In a
system of interrelating parts, this includes aspects such
as the topological structure defining the "social network" of the system.
It has long been understood that social and economic activity occur on networks, and the network structure
can play a crucial role in determining the outcome of
such social or economic activities [41]. On top of this,
in open socio-technical systems, such networks are
often dynamic in nature, as individuals arrive and
leave, and the system's topology forms and reforms on
an ongoing basis. While theoretical and empirical
research have led to a substantial increase in our
understanding of the effect of networks on large scale
systems, extending our understanding to temporal
networks remains a hot topic with numerous open
challenges [23].

It has long been understood that
network structure can play a crucial
role in determining the outcome of
social or economic activities.
Socially-sensitive systems will need to be aware of,
and act in accordance with the impact of changing system organization. This will impact economic outcomes,
efficiency (in terms of information propagation, for example), distribution of control, and the presence (or not)
of emergent behavior.
Where such a system adapts its own organization,
the process is referred to as self-organization. Selforganization is often found in large-scale, complex, and

IEEE TECHNOLOGY AND SOCIETY MAGAZINE

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - September 2017