IEEE Technology and Society Magazine - Winter 2014 - 74

is to use the increased connectivity
provided by wireless communication
technologies, the increased computing
power, and the memory capacity of
embedded devices in order to implement autonomous behavior, which
can support the citizen (user) in everyday life. The concept of "citizen," and
not simply "user," is deliberately introduced here to suggest that the human
capacity to maintain autonomy and
control in a world of pervasive humantechnological networking should be
considered as an essential part of our
ethical and legal endowment and entitlement in IoT. But one should realize
that the "things" in an IoT scenario
are new entities (new ontologies).
These new entities are endowed with
identity, connectivity, intelligence,
and agency with and through which
relationships ("contracts"), different
from the customary, emerge between
the "things": subjects with objects and
objects with objects. This observation
poses many questions of a social, ethical, and legal nature.
A classical IoT use case is smart
home care, where sensors collect
information (e.g., body temperature,
blood pressure) about an elderly person and other home sensors and actuators regulate the environment (e.g.,
temperature, humidity) of the house
or provide medicines in order to optimize the condition of the person.
This use case already identifies some of the potential issues in
the relation between the citizen and
the developing IoT vision. What
type of policies should regulate
the autonomous behavior of the
"things," including the delegation
of human autonomy to IoT objects?
For example, could a smart home
provide medicines to the elderly
person in case of need? And who
decides which medicines? What are
the threats and acceptable risks for
the elderly person? Who is accountable for the proper functioning of the
system? Who is liable in case of serious damage or death of the elderly
person because of an unpredictable
functioning of the smart home or lack
of assistance if the smart home did
74

not act? In addition, how is personal
data privacy ensured and restricted to
provide only the relevant information
required by the medical personnel?
Who decides what "relevant" means?
Who is responsible for providing
assistance to the elderly person?
Questions also remain about the
threats of delegating fundamental
aspects of humanness, such as care
for the weak and the elderly, to IoT
smart devices, which may not have
the same cognitive capabilities to
replace basic human skills for caring (see [3]). Beyond the smart home
care use case, the increased pervasiveness of IoT in our everyday life
calls for a more structured approach
for the development and deployment
of IoT technologies and their interaction with citizens. This interaction is
at least expected to be secure, respectful of one's privacy preferences,
trustworthy, and accountable. In a
continuously changing and unpredictable environment from the social,
technological, organizational, and
regulatory points of view, such expectations may be challenging to fulfill.
In this article we define trust (definition adapted from [4]) as the level
of confidence with which an entity
can ensure to another entity or entities
specific services tailored for given
contexts and quality (fitness for purpose and reliability). Moreover, trust
is also related to the user's belief that
an entity or system functions normally and delivers specified services
on the basis of evidence, current and
past behavior, design, types of certification of the device, user's personal
knowledge, perceptions, and previous
experiences. Both the subject-related
and machine-related trust levels
can be measured using appropriate
qualitative and quantitative methods
(which is also referred to as trustworthiness). In IoT, these different
interactions are co-dependent and
therefore can be hardly separated.
The importance of trust in human
machine interaction systems or automated services has been highlighted
in [5], where the authors describe
how trust can support sustainability

of social networks. In the specific
case of location-based social networks, the authors discuss in [6] the
impact of this type of networking in
the trust assigned to the relationships.
Our vision in this paper is that agency,
as the individual ability to intervene
and tailor the system, is a crucial element in building trust in IoT technologies. Following up on this vision, we
will first address the issue of agency,
namely the individual capability to
adopt free decisions, as a relevant
driver in building trusted human-IoT
relations, and how agency should be
embedded in digital systems. Then we
present the main challenges posed by
existing approaches to implement this
vision. We show then our proposal for
a model-based approach that realizes
the agency concept, including a prototype implementation.

Promoting Trust In The
Human-IoT Relationship:
Enhancing Agency through
"Rights in Design"
The preservation of the human
capability to freely act and make
choices has been shown as a major
factual and normative concern
among the different requirements
for a sustainable governance of IoT.
Indeed, a normative framework
for a viable vision of IoT requires
a renewed proactive attention -
and moral commitment - towards
human agency, namely towards
a full concept of humanness and
towards preserving the features of
human action. Agency represents
a broader concept than autonomy,
encompassing not only the specific
cognitive and free will attitudes of
an autonomous self, but the wider
range of acts and activities connected to human life [7].
The preservation and cultivation
of human agency in IoT is especially
at stake as the embedded (and often
embodied) pervasiveness, miniaturization, distributed, and networked character of connections can impair the
human ability to control the system
while being a part of it. This implies
creating the conditions for a dynamic

IEEE TECHNOLOGY AND SOCIETY MAGAZINE

WINTER 2014

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - Winter 2014