IEEE Technology and Society Magazine - June 2017 - 56

with robots due to an increased degree of anthropomorphism [8], [54].
The impact of gendering robots on society is likely
to both entrench and subvert traditional gender norms
[15]. To manage the complexity of the design challenge, designers may single out certain qualities and
neglect others, building simplistic characters with stereotyped forms and behaviors [55]. Designers may
hark back to an earlier era to which they feel nostalgic,
in a process known as "retro-tech." As a consequence
they may apply dated stereotyped gender roles to their
designs that do not fit with current social norms [2].
We can see this in robots developed in both Japan and
in the U.S.A.

In general, robot designers show
little awareness of the contingent
relationship between gender and
human bodies.

The gendering of robots, and society's willingness to
alter its behavior and perceptions around robots, is likely to reinforce gendered stereotypes. As we willingly
adapt ourselves to our robotic creations, a gendered
humanoid robot could become the ultimate gendered
being, a perfect woman or man that an imperfect,
flawed human being is unable to compete with. Consequently, humans may feel they need to live up to the
expectations created by humanoid robots.
Gendering humanoid robots will also change our
social and cultural ideas of gender in human bodies.
Assigning gender to robot bodies reinforces the idea
that gender is an assignation, a set of attributes [4].
Potentially the gender of robots can be changed, as in
the case of a Japanese designer who created a female
humanoid robot. When asked if he could make a male
robot, he simply changed the wig and clothing. For
robots, gendered attributes are easily interchangeable. Users may wish to change a robot's gender
dynamically in different contexts. If humans identify
with humanoid robots, and see that robot gender is
interchangeable, we may start to question why human
gender is fixed. This would challenge dominant social
norms about the immutability of human gender, which
could lead to a greater fluidity of gender, as argued by
cyberfeminists, and encourage society to adopt a more
flexible view.

The data mining of gender and assignment of attributes may entrench stereotypes, or present opportunities for greater gender diversity. Using novel algorithm
data structures, robots may learn from the vast repositories of information communicated throughout
human history. If large datasets contain stereotypes,
the robot will ultimately learn and apply these in their
social interactions. Researchers at Microsoft asked a
system trained on Google News articles to complete
the following analogy: "man is to Computer Programmer as Woman is to x," the response came back:
"homemaker" [56]. Robots learning from these sources may provide an uncomfortable mirror of our stereotyped views of men and women. Conversely, if
machines "learn" by engaging with thinking in the
human world, they may move beyond their limited, programmed role and embrace identities more along the
full continuum of gender. Thus a fully autonomous
robot may behave in a different way than what is
expected by its gendered appearance.
Here, at the cutting edge of research, robots are
beginning to learn and adapt to humans. By observing
human speech and language, robots can process information and respond. According to Cuayáhuitl, current
research includes locating the human speaker, their language and intended meaning, and creating a natural
dialogue for HRI [57]. Robots may learn by observing
and mimicking the physical movements of humans performing tasks [58], which may also be influenced by
gender stereotypes.
Giving a robot realistic human attributes, such as
gender, allows humans to anthropomorphize a robot
to the extent that they may believe they can have an
emotional relationship with it [54]. In the film Ex
Machina, the human Caleb is transfixed with the robot
Ava. To his detriment, he mistakenly believes that they
form an emotional bond. This willingness of humans
to accept humanoid robots as equals is facilitated by
the pervasive use of machine metaphors to understand the human body. These metaphors assist HRI by
breaking down the rigid boundary between human
and machine, allowing us to see ourselves as closer to
robots [59]-[62]. As Richardson notes, robot designers at
M.I.T. already perceive humans as a type of machine [48].
Realistic humanoid robots have the potential to disrupt social interaction and create less empathic relationships, by offering a poor substitute for human
connection. Richardson argues that the development of
sex robots could reduce the empathy and reciprocity of
human-human relationships [21]. While a humanoid
robot may present an appealing solution for social problems such as loneliness and the need for intimacy [63],
it ultimately provides only a mirror, allowing humans to
indulge in narcissistic relationships with themselves

IEEE Technology and Society Magazine

∕

june 2017

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - June 2017