IEEE Robotics & Automation Magazine - March 2011 - 40

(b)

(a)

(c)

(d)

Figure 1. Robotic applications of (a), (b) service and (c), (d) combat robots. (a) Childcare robot PaPeRo [32], [73]. [Photo courtesy
of NEC Corporation.] (b) Paro therapeutic robot [89]. [Photo courtesy of AIST, Japan.] (c) MQ-9 Reaper Hunter/Killer UAV by General
Atomics Aeronautical Systems [33]. (d) Special weapons observation reconnaissance detection system (SWORDS) by Foster-Miller [42].
[Photo courtesy of Foster-Miller.]

robots may become sentient machines [51], which we
would no longer be allowed to enslave [75], or that we
may create robots capable of annihilating mankind [17].
For a discussion on these issues, we refer the reader to
[56], [75], and [17].
Who or What Is Responsible
When Robots Harm?
Veruggio [100], [102] dates the beginning of "roboethics"
from two events. One was the Fukuoka World Robot
Declaration, wherein it was stated that "next generation
robots will contribute to the realization of a safe and

(a)

(b)

(d)

(c)

(e)

Figure 2. (a) The responsibility-ascription problem, i.e., the
problem of assigning responsibility to the manufacturer, designer,
owner, or user of the machine when use of this machine led to
an armful event is a yet largely open issue. (b) People tend to
blame the robots because they falsely attribute them with moral
agency [29]. (c) People blame the machine even if they recognize
the machine's lack of free will and lack of intentionality [28]. (d)
Many ethicists argue that we should to some extent hold the
engineers (the creators of the malfunctioning robots) responsible
[60]. (e) To do so, we should use existing the legal principles, or
create new ones, if necessary [13].

40

*

IEEE ROBOTICS & AUTOMATION MAGAZINE

*

MARCH 2011

peaceful society." The other was the roboethics road map
[101], which sought to promote a cross-cultural discussion
among scientists to monitor the effects of robotics technologies currently in use. More recently, an initial sketch of the
code of ethics for the robotic community has been proposed
[43]. This code offers general guidelines for ethical behavior.
For example, the code reminds engineers that they may be
held responsible for the actions of artificial creatures that
they have helped to design. Along similar lines, Murphy and
Woods [70] propose to rephrase the famous Asimov's laws,
which they view as robot centric, in such a way as to remind
robotics researchers and developers of their professional
responsibilities. For example, the first law was replaced with
"A human may not deploy a robot without the human-
robot work system meeting the highest legal and professional standards of safety and ethics" [73, p. 19].
All the above implicates the responsibility ascription
problem [69]: the problem of assigning responsibility to the
manufacturer, designer, owner, or user of the robot or to the
robot itself when using a robot leads to a harmful event.
From a philosophical perspective, it is generally agreed that
robots cannot themselves be held morally responsible [9],
[25], [38] (although a few oppose this [95]) because computers as we conceive them today do not have intentionality
[28]. From a psychological perspective, however, it remains
an open question whether people include robots as an additional agent in the ascription of moral responsibility.
Who or what is responsible when robots harm (Figure 2)?
Matthias [62] provides a seemingly simple answer. He
argues that, in most cases, no one can be held accountable
for the robotic failures. Matthias argues that with the
advance of programming techniques (e.g., neural networks,
evolutionary computation) that equip the agent with the
ability to learn and, hence, to depart from its original
Table of Contents for the Digital Edition of IEEE Robotics & Automation Magazine - March 2011
