IEEE Technology and Society Magazine - March 2020 - 55

and responsibility between the two. That balance must
take into account the behavior and skills of each party,
and it is mediated by the variation and intensity of communication between the two parties [16].
Since there is a behavioral asymmetry between
human and robot, as much control and responsibility
for the interaction as possible should remain with the
human. The autonomy of the robot should be constrained within boundaries.
Because the emotional intensity of stimuli is overwhelming for the autistic child, the child seeks to get
control by various means, whether by stimming, or
shutting down [8]. The child may seek to hide, or lash
out in an attempt to manage and reduce the emotional
intensity. The child needs more control over his environment. If the robot is to be of help it should reduce
the sensory demands on the child. And the decision as
to whether to interact with the robot must be the
child's. If the robot takes autonomy from the child, the
child may become more distressed. The role of the
robot should be to empower the child by increasing the
child's autonomy.
Any wresting of autonomy from the human in the situation needs to be ethically justified. The presence of a
robot needs to be liberating, creating a sense of safety.
The robot should not replace any activity the child could
do for himself. The robots in the DREAM project, implementing ABA, may decrease safety and increase anxiety: not a good idea.

Community
An interaction with a social robot takes place within a
social environment and alters that environment. The
presence of a robot may take the focus off the child.
Richardson et al.'s [20] ethics addressed a set of disparate stakeholders, principally experts and not the community that revolves around the child. Working with an
autistic child, the robot is part of a team, mediating the
effects of various relationships between the child and
his carers.
One danger in working with a robot is that we shift
from a community focus to an institutional focus. The
child becomes the subject of experts and institutions.
In considering ACTIVE ethics, we shift from a set of
disconnected stakeholders to a community of connections that surround the child at home, at school, and
at play, and which create a safe environment. The
DREAM project's quantitative survey asked participants "Is it ethically acceptable that social robots are
used in healthcare?" Eighty-five percent agreed with
this vague question. The question should be not
whether but why. While robots trundling around a hospital delivering food and drugs may be fine [19], the
use of robots to replace therapy will be socially
MARCH 2020

∕

isolating. The autistic child needs social connections
provided in a safe environment.
The focus of the community living, teaching, and
working with the autistic child should be on the child,
not the technology. As such the robot comes into the
community as a guest that needs to be fitted into the
team's purpose.
The ethics of community is one of individual respect,
of deep relationships between people. Community is a
place of acceptance and safety. A robot ethics that
dehumanizes, as the DREAM project robot seems to,
degrades community.
Ill-thought-out, insensitive deployment of social
robots will only succeed in dehumanizing and disconnecting the community. Deployment of a social robot
requires a well-connected community with coherence
and good boundaries, such that the presence of a robot
can be absorbed. The community should agree on -
and understand - the purpose of the robot intervention. That purpose cannot be imposed but must be
negotiated. This also requires that the robot's purpose
and limitations are understood. This requires our third
ethical focus of ACTIVE ethics, on transparency.

Transparency
The use of social robots cannot be conducted ethically
without transparency. First there must be transparency
as to the goal and purpose of robot deployment. In the
DREAM project this is not made clear, because the
underlying goal of ABA is not explored and articulated.
Hence the controlling element of robot deployment
remains hidden. Second there must be transparency
concerning the nature and limitations of the robot. The
NAO robots in the DREAM project are programmable
machines. Pretending that a robot is not a machine is
as unethical as treating a human as a machine.
Transparency requires the education of the community with sufficient knowledge to understand the
operation and programming of the robot. Presenting
the robot as if it's a person may create expectations
that it can do more physically and socially than is
actually the case.
NAO, the robot used in the DREAM project, is a programmable robot that can carry out sets of routine
maneuvers, such as dances, and act as an educational
tool. It is not empathic, emotional, or really engaged
with the human. It is truly robotic.
The use of radical behaviorism in robot enhanced
therapy involves social programming of the child by
repetition. The child sits in front of the NAO robot
repeating what the robot says. The aim of the programming of the NAO robot in the DREAM project is not just
to replace the human therapist in providing input, but
also to sense the state of the child, to evaluate the

IEEE TECHNOLOGY AND SOCIETY MAGAZINE

IEEE Technology and Society Magazine - March 2020

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