IEEE Technology and Society Magazine - March 2020 - 58

developing empathy we can consider how a robot
might support the child rather than using the robot to
imprison the child in perceived social conventions.

DREAM Project Ethics
Using the paradigm of radical behaviorism, the DREAM
project sought to isolate the autistic child in a smart
room, with a robot controlling his behavior, whereby
smart detectors embedded in the room would monitor
the child's behavior and reaction to the robot and feed
this data to the NAO robot which would change its
instructions to the child accordingly. Such imprisonment and forced behavioral change is clearly inhuman
and unethical.
While the discussion of ACTIVE ethics in this paper
has picked this up, this ethical aspect was not de tected by the ethics function in the DREAM project.
Why not?
Firstly the underlying assumption about the nature of
humans and the nature of autism were not adequately
questioned. Baron-Cohen's model of ASD [9], which
underpins the perception of the DREAM project, presents only one view of the nature of autism and the way
the autistic mind works. But replacing a materialism
model with a social model that views autism as merely a
social construct is equally flawed. It is not only a matter
of social perception. There is real physical dysfunction
in the brain. And the social problems that can devastate
families, destroy marriages, and isolate the child are
very real and require therapeutic intervention. But the
robotic procedure of ABA is not what is needed.
Furthermore, viewing autism as a lack of empathy and
the child as robot can be used as an excuse for justifying
the use of robots. I have already suggested that there is a
vast mismatch between the behavioral repertoire of an
autistic child and a NAO robot. The child is anything but
unempathic and rather may be viewed as oversensitive to
emotional stimuli and unable to detach. If we fail to
understand the premise and purpose of the behavior we
are programming into the robot, we open ourselves up to
ethical failure. Additionally, an interest in technology
does not mean that the child is especially averse to
socializing or should be isolated with robots. The autistic
child wants friends as much as anyone else but lacks the
brain pathways to make that happen. He oscillates
between social stress and severe loneliness [17]
Secondly, the ethical procedures of the DREAM project fail to engage with the central protagonist of the
drama, the autistic child.
The investigation of ethics in the DREAM project [20]
involved a wide range of stakeholders including clinicians, parent, autism and robotic experts, school heads,
and technology professionals. The field work involved
interviews with parents, with experts in the technology

and healthcare practitioners. Academic ethics scholars
were packed into workshops to provide a range of theoretical and philosophical insights. Public meetings,
forms of so-called deliberative democracy, involved
members of the public considering very generic aspects
of using robots in healthcare following presentations
from experts. Such meetings bore little relevance to the
practice and purpose of the therapeutic robots. A quantitative survey of parents, teachers, and therapists
agreed in general terms to the use of social robots.
While 20 robot ethics workshops were conducted, only
two autistic children were interviewed [20]. The children
were almost an afterthought in the ethics of the project;
those to whom the greatest responsibility is owed are
almost left out. The robot ethics of the DREAM project
should have been the ethics of the client [14], which in
this case is the autistic child. Rather than attending workshops on robot ethics, substantial time should have been
spent with autistic children, getting to understanding
them, developing empathy and creating connection.
Thirdly, the focus on technology and the ethics of technology distracts from a focus on the ethics of the processes and procedures that the technology is supporting or
implementing. In this case the origins, nature and ethics
of Applied Behavioral Analysis remained unexplored.
There was no attempt to understand its historical roots, to
peel back the philosophy and view of human life that
underpins it, to subject it to ethical scrutiny, nor to consider its outcomes and results. Some autistic children who
have been exposed to ABA suffer from post-traumatic
stress disorder. One study suggested this was the case in
for 50% of females who had undergone ABA [10].
We cannot generate an ethics of technology by
addressing generic characteristics of the technology. It
depends what is done with the technology and by whom.
A robot that dispenses medicine according to pharmacy
staff orders [19] supports human flourishing; a robot that
pins patients down for an injection does not. Artificial
Intelligence that makes fast work of pattern recognition
in diagnosis of diseases such as eye disease is of great
human benefit. AI that makes decisions on treatments,
on who lives and dies, and overrides, or is considered
more authoritative than humans, is not.
ACTIVE ethics has achieved what the traditional
approach of the DREAM project could not in exposing
key issues and highlighting the debatable ethical practices of ABA therapy and robot assisted therapy. It has
done this by taking a human view and positioning the
autistic child at the center of a community rather than
deploying a political and institutional ethics.

Way Forward for Robots and Autism
Although the use of robots to force operant conditioning
is inhuman, that does not mean robots cannot have a

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