21robots everywhere
very strange. It seems as if we are dealing witha kind of uncanny val-
ley, in the sense that the language seems real, but not the social inter-
action. e conversation between these two computers is, however,
not at all creepy, but is rather humorous.
Social interaction between people concerns not just verbal infor-
mation, but, more importantly, nonverbal communication as well;
think about posture or emotions that can be read from facial expres-
sions. “Aective computing” deals with this area of the human–
machine interaction process. According to one of the founders of
this eld, Rosalind Picard (1995) of MIT, we are dealing with “com-
puting that relates to, arises from, or inuences emotions” (p. 1). e
goal is that computers learn to recognize human emotions and learn
to adapt their behavior on that basis. To this end, aective comput-
ing analyzes aspects such as intonation of the voice, gestures that
people make, bodily posture, and facial expression. For example, the
Dutch company Noldus has developed FaceReader, which is used
regularly by marketing researchers. is technology uses the Facial
Action Coding System (FACS). is was developed by Paul Eckman, a
renowned psychologist, who as far back as the 1970s suggested that
there are six basic human emotions—anger, disgust, fear, happiness,
sadness, and surprise—all of which can be read from the face within
a millisecond. is coding can also be used so that avatars, softbots,
or real robots can show emotions. It is expected that the user friend-
liness, and thus the acceptance, of such technologies will increase
(Picard & Klein, 2002).
1.2.4.6 Artificial Morality e question of whether not only social
behavior but also moral behavior can be programmed into comput-
ers is currently being discussed. is is quite clearly a very recent
scientic eld. At the beginning of this introductory chapter,
there was a reference to the three Asimov ethical laws that Asimov
robots are supposed to comply with. Especially in the eld of mili-
tary robots, there is reection on the use of robots, which should
behave according to international humanitarian law, as dened in
the Geneva Convention. Ronald Arkin (2009) assumes that it is
possible to develop robots that can make better decisions under
combat conditions than human soldiers. He proposes not only that
AI is independent of emotions, as it is only based on logic, but at
22 Just ordinAry robots
the same time that the use of such moral machinery will limit the
number of military casualties to a minimum.
Wallach and Allen (2009) are also in favor of the development of
so-called articial moral agents. ey argue that “today’s information
systems are approaching a level of complexity that … requires the sys-
tems themselves to make moral decisions—to be programmed with
‘ethical subroutines’ to borrow a phrase from Star Trek” (2009, p. 4).
ey suggest that people are unable to oversee all the consequences
of very complex interacting software systems, which could therefore
possibly lead to catastrophes. For example, the automation of nan-
cial transactions, or robotic trading, led to the 2010 Flash Crash.
OnMay 6th, the Dow Jones Industrial Average rst plunged about
1000 points within minutes—about U.S. $1 trillion in market value
vanished—and then recovered those losses in less than 3minutes. It
is telling that there still is no consensus on the exact root of the 2010
Flash Crash (Steiner, 2012, p. 4). As more and more of our world is
placed under the control of multiple linked algorithms, it becomes
more dicult to supervise or gain insight into who or what is causing
undesired catastrophic events. In order to identify such catastrophes at
an early stage and to be able to prevent them, Wallach and Allen hold
that software bots should exhibit ethical behavior.
However, just as a robot with social intelligence is an engineering
dream, a machine that exhibits full moral agency is also a futuristic
vision. e idea of a moral robot functions as a spot on the horizon
that may help to dene pathways from current to more sophisticated
technologies. James Moor (2006) denes four categories of machine
ethics and/or articial moral agents: (1) ethical impact agents,
(2) implicit ethical agents, (3) explicit ethical agents, and (4) full ethi-
cal agents. Since, by its nature, computing technology is normative, all
robots can be considered ethical impact agents that can be evaluated
for their ethical consequences. At this rst level, machine e thics is very
close to mainstream computer ethics. At the second level are implicit
ethical agents: machines that are designed so that they implicitly pro-
mote ethical behavior or at least do not have negative ethical eects.
Over the past decade, the interest in addressing values in design
has grown. In particular, value sensitive design has become a widely
accepted approach to design (information) technology that accounts
for human values in a comprehensive manner throughout the design
23robots everywhere
process (Friedman, Kahn, & Borning, 2006). e next two levels con-
cern the engineering challenge of incorporating ethical decision mak-
ing into machines. Until this moment, it is not yet clear to what extent
this might be feasible. At the third level come explicit ethical agents,
which are machines that have human morality encoded in their soft-
ware. Finally, Moor wonders whether a machine, like an average adult
human, can be a full ethical agent that can make explicit ethical judg-
ments and is generally competent to reasonably justify them.
1.2.5 Networked Robots and Human-Based Computing
In the analysis of robots, all eyes are often on the robot itself. However,
the performance of many robots depends heavily on the support of
other technologies plus human intelligence. e term networked robots
indicates that these robots are supported by various information net-
works, without which they could not function. Military drones above
Afghanistan, for instance, make use of 32 global positioning system
(GPS) satellites, of which only 4 would be necessary for an unmanned
aircraft to determine where it is located. In addition, smart machines
often need human intelligence to become really clever systems. As
Floridi (2014) points out, “[S]ometimes our ICTs need to understand
and interpret what is happening, so they need semantic engines like
us to do the job” (p. 146). is recent trend is known as human-based
computing. We have already described how the Cleverbot generates
answers based on earlier answers from people to similar questions
that can be found on the Internet. In a similar vein, Google Translate
searches through millions of documents that are translated by human
translators to come up with a translation within a fraction of a second.
Expert systems are dependent on large databases for their operation,
and the Internet now provides a great deal of information in many
areas, from shopping behavior to human faces. In our exploration of
the new robotics, we therefore continue to pay attention to the net-
works and human eort behind the robot. If we did not do this, the
signicance of the modern robotics society would be incomprehensible.
e popular, but incomplete, image of the robot as an inde-
pendent and self-sucient machine is supported by contemporary
visions. But it is probably also a relic of past thinking about the
future. In the 1950s and 1960s of the last century, the robot was
24 Just ordinAry robots
often depicted as an independently moving machine with legs and
sensors. At the end of the 1960s, futurologists Kahn and Wiener
(1967, pp. 86–98) thought that up to the year 2000 the inuence of
computers would primarily be through the automation of mechani-
cal machines. is was true for production, and also, for example,
for domestic work. Around the year 1984, the English mechani-
cal engineer ring already foresaw a robot “having no more feel-
ing than a car, but having a memory for instructions and a limited
degree of instructed or built-in adaptability according to the position
in which it nds various types of objects. It will operate other more
specialized machines, for example, the vacuum cleaner or clothes-
washing machine” (quoted in Kahn & Wiener, 1967, p. 94). Kahn
and Wiener saw the automation of information processes as the next,
more dicult, step. ey quote Lipetz (1966), who thought that by
means of these developments “the geographical boundaries of tra-
ditional information storage and retrieval systems are beginning to
evaporate. In their place are beginning to emerge vast networks of
compatible communication devices linking users with many special-
ized and overlapping collections” (quoted in Kahn & Wiener, 1967,
pp. 95–96).
In recent decades, however, the automation of machines and infor-
mation processes has occurred simultaneously and often go hand in
hand. is may also be explained from the perspective of the conver-
gence of IT with other technologies. e eect of IT on various other
technical elds—also known as digitizing—is often referred to by the
term convergence. e automation of all kinds of production processes
requires the convergence of mechanics and electronics, or mecha-
tronics, which is the basis of the industrial robots. e emergence of
the Internet depended on the convergence of information and com-
munication technologies, labeled ICT. e high expectations of the
new robotics are based on the convergence of the Internet and robot-
ics that is expected during the coming years. e development of the
new robotics is thus carried by the Internet, and this also continues
to change it.
During the past decade, the Internet has penetrated the whole of
society. ree technology trends are responsible for this (Van’t Hof,
Van Est, & Daemen, 2011, pp. 128–130). e rst trend is the rise
of digital devices in public spaces, from ATM terminals, cameras,
25robots everywhere
gates, and navigation systems to smartphones. Digital convergence
is the second trend. is trend means that the networks behind these
devices are increasingly linked to the Internet—the mother of all net-
works. e third trend is that in recent years the Internet has become
available in ever more places, especially through the smartphone.
Seen from this perspective, service robots are a new kind of smart
device that will populate our world. On the one hand, the new robot-
ics is based on existing networks. Conversely, it thereby changes the
nature of these networks. is is clearly expressed by the vision of
robotics as developed in the United States in From Internet to Robotics
(Christensen, 2009). is indicates that robotics builds on and uses
the existing ICT infrastructure. But the message is that robotics is a
further technical development of the existing information networks.
e worldwide web has been extended with robotics, giving the
Internet “senses and hands and feet.”
1.3 Seen Socially
In this book, we want to investigate the social signicance of robotics
in the years to come. We do this by studying robotics developments
in ve dierent areas, surveying the home, long-term health care,
police and private drones in cities, trac, and the army. e central
approach is that the use of robotics will aect these elds of applica-
tions and social practices in many ways. We will describe four key
characteristics of modern service or social robots that produce various
social and ethical issues and thus raise questions of social acceptance:
(1) robots as IT, (2) the lifelike appearance of robots (body and brain),
(3) the level of autonomy of robots, and (4) robotization as rationaliza-
tion. is gives the reader a number of thematic tools for reading the
following ve chapters.
1.3.1 Information Technology
Robots are IT. is means that social issues such as privacy, cyber
security, the digital divide (access to technology, computer skills), algo-
rithmic transparency, and data ownership also play a role in robotics.
e fact that within robotics great attention is given to improving the
interface between machines and humans brings new questions with it,
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.