IEEE Technology and Society Magazine - March 2020 - 54

One aspect of autistic behavior can be repetitive
movements. These may be self-damaging such as the
child hitting his head against a wall, or poking himself
in the eye. Less damaging is rocking or humming. This
behavior, known as stimming, is the child's attempt to
gain control of the overwhelming emotions and stimulation affecting the brain [23].
While the external behavior may seem disengaged,
the autistic mind is often in overdrive, finding it hard to
stay on one thing, to manage the cascade of stimuli.
For children less on the autistic spectrum, concentrating to an obsessive extent on one subject, whether car
number plates, train numbers, or chemical structures,
enables a calming effect, and helps block out the overwhelming stimuli [2].
The autistic response is a retreat from social interaction into technology because the social demands are
overwhelming and confusing. This withdrawal, which
may involve lashing out, is a protective mechanism to
create safety in an unsafe world.

Robot use in increasing a child's
autonomy should involve the child
telling the robot what to do, not the
robot telling the child what to do.

The child is actually eager for social contact and
would like to be part of a social group [21]. Isolating a
child with a robot will only increase social frustration
and inhibit social connection. The child wishes he could
play with other children, but finds this difficult.
The autistic child and the robot are not behaviorally
equivalent. The variety and degree of emotional responsiveness of the autistic child vastly outstrips the robot,
which raises the question as to whether there is any
point at all in using robots in autism therapy beyond the
rhetoric of replacing human workers.
There may well be. Robots may be used to increase
the child's autonomy. They may help the child gain control within a therapeutic setting and manage the social
complexity of the environment. This will involve the
child telling the robot what to do, not the robot telling
the child what to do.

Towards an ACTIVE Ethics of Robot
Enhanced Therapy
Richardson et al.'s [20] report on the DREAM project
describes the ethics research behind the project but

reaches few conclusions about the ethics of robot
involvement in autism. Rather it only addresses the limited brief of the DREAM project. I would suggest that it is
ethically questionable to make the robot the focus of
research project activity, as occurs in the DREAM project, rather than the user. Indeed it is questionable whether it is ethically justifiable to use robots as human
replacements in situations where social interaction is
critical. Neither is it ethical to subordinate the child to
the technology. The ethics of DREAM, while addressing
classical ethical issues such as privacy, failed to consider
the key needs of the child or treat the child with enough
humanity. A majority of the fieldwork did not address the
children, but the academics, professionals, and research
groups around them. The children became invisible.
In an attempt to be deliberately democratic, inclusive, and scientific, the ethics fail and the children's
needs are not served. Rather the children are subservient to the needs of the health authorities, clinicians,
and research agencies. This is because the underlying
therapy seems to be directed to the needs of the societal stakeholders and not the children.
The point here, which will be addressed further, is
that we cannot treat technological ethics as separate
from societal ethics. The behavior of the technology is
derived from its application and purpose in society, and
does not operate in isolation by itself.
In order to gain a foothold in the ethics of social
robots we must shift the focus from the technology and
the technologist to the subject, the patient, the vulnerable person for whom we consider robots to be a possible viable intervention. It is not a case of robot ethics
but Asperger's or ASD children ethics.
To illustrate such a shift, I will apply the ACTIVE ethics framework [15] to ASD robot-assisted therapy and
hence identify the ethical positioning of robots in
autism therapy. The result of the application of ACTIVE
ethics is that we can place robot usage such that it delivers benefit for the autistic child.
ACTIVE Ethics considers six ethical areas: Autonomy, Community, Transparency, Identity, Value, and
Empathy.

Autonomy
The DREAM description of work [20] refers to therapy
robots needing to become more autonomous. The
intention would be that the robot replaces the therapist,
manages the child, and is effectively in charge.
Autonomy concerns the locating of control. The aim
is to locate control with the robot. But aims to locate
control with the robot, in other words to attain total
autonomy, are not only impossible but dangerous in the
case of social robots. Managing autonomy in a robothuman relationship requires the balancing of activity

IEEE TECHNOLOGY AND SOCIETY MAGAZINE

∕

MARCH 2020

IEEE Technology and Society Magazine - March 2020

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - March 2020