IEEE Systems, Man and Cybernetics Magazine - July 2019 - 53

to job security and human safety in the past. However,
benefits from industrial robotics outweighed the problems they imposed, and society has been able to see that.
This visibility was made possible by the contributions of
many scientists, engineers, and policy makers.
We are now at a similar point in time when we need to
investigate the reasons for resistance to autonomous
systems and evaluate their impact on our society. As an
integral part of this investigation, we must explore the
relationship of trust between humans and autonomous
systems (or, more specifically, autonomous robots) and
understand the dynamics of trust. Possibly, we need to formalize how and when trust is gained by a nonautonomous entity, whether it is a toy, a consumer product
like a mobile phone, or a crucial life-saving device. We
must also explore the process needed to convince society
of the benefits of having an Autonomy of Things.
In general, trust is usually driven by the willingness of
an agent, the trustor, to be vulnerable to the decisions and
actions of another agent, the trustee, based on the belief
that the other will execute a particular action irrespective
of the ability of the trustor to monitor or control the trustee. We can reasonably assume that the complexity of a
manual task has a strong correlation with the possibility of
its being delegated to an autonomous entity. This delegation also correlates to increased mental and cognitive load
as well as associated risks. This leads to a logical conclusion that the epitome of trust in an autonomous system
will be the deliberate delegation of any task by the trustor
to the autonomous trustee (see Figure 1).
Difficulties in Measuring Trust
We are in dire need of understanding the process of trust
in the human mind through measurable, and quantifiable,
biological signals. These biosignals, or biomarkers,

Transparency and Predictability
Trust depends on these attributes
during the formation process and
on an ongoing basis.

including movements, act as a proxy for trust in the
human mind, and they are essential in this exploration, as
we cannot measure our minds. To date, trust can be identified only by asking a human explicitly about another
human or a nonhuman entity in a particular scenario.
This is not suitable, and a true response is not possible to
achieve, in many cases.
However, humans have a very complex chemistry in
every mind-related process, and trust is no exception.
Often humans cannot conclusively answer whether they
trust someone or something. This is because trust is a multifaceted topic and varies in nature across situations, use
cases, and environments. Trust in another human being
can also be in relation to a task to which a particular
human being is assigned or can, in rare contexts, be universal. For example, person A may trust person or entity B
with task X but not with task Y. A reasonable person will
trust a mobile phone to make a phone call to family but
will not trust it to work as a tourniquet to stop blood loss,
even though he/she can use the edge of a thin mobile
phone to press an artery.
Finding and associating biomarkers of the trusting
process is a very difficult task if we have no prior knowledge about the situation. This a priori knowledge about
any situation is also subjective and may not be accurate
if collected from human participants via a questionnaire.
It imposes a challenge in the data collection process of
any trust-related experiment by compromising the quality of data.
This particular problem gets even more complicated if a
human participant is asked directly, in the form of a questionnaire, whether he or she is trusting in a particular scenario. The answer is dependent on the relationship
between participant and investigator, the investigator's
approach, the participant's geo-cultural background, and

Prospective Biomarkers for
Objective Measurements
These include electroencephalogram, heart rate, perspiration,
eye movement, electromyogram, breathing rate, and
breathing sounds.
Subjective Nature
Depends on context, age, culture,
environment, education, and profession,
among other factors.

Trust

Control Is Important in
Human-Machine Trust
Due to the historical nature of
exposure to machines, having
control is important for humans
in the process of trust formation.

Autonomy of Things
Trust will play the pivotal role in the early
days of the Autonomy of Things.

Defined by Willingness of Someone
Trust is created in the vulnerability
of a willing party.
Figure 1. Trust is a multifaceted concept that requires consideration from numerous perspectives.

Ju ly 2019

IEEE SYSTEMS, MAN, & CYBERNETICS MAGAZINE

IEEE Systems, Man and Cybernetics Magazine - July 2019

Table of Contents for the Digital Edition of IEEE Systems, Man and Cybernetics Magazine - July 2019