Rule based fuzzy cognitive maps and natural language processing in machine ethics
| DOI | https://doi.org/10.1108/JICES-10-2015-0034 |
| Published date | 08 August 2016 |
| Date | 08 August 2016 |
| Pages | 231-253 |
| Author | Rollin M. Omari,Masoud Mohammadian |
| Subject Matter | Information & knowledge management,Information management & governance,Information & communications technology |
Rule based fuzzy cognitive maps
and natural language processing
in machine ethics
Rollin M. Omari
School of Computer Science, Australian National University, Canberra,
Australia, and
Masoud Mohammadian
Faculty of Business, Government and Law, University of Canberra,
Canberra, Australia
Abstract
Purpose – The developing academic eld of machine ethics seeks to make articial agents safer as they
become more pervasive throughout society. In contrast to computer ethics, machine ethics is concerned with
the behavior of machines toward human users and other machines. This study aims to use an action-based
ethical theory founded on the combinational aspects of deontological and teleological theories of ethics in the
construction of an articial moral agent (AMA).
Design/methodology/approach – The decision results derived by the AMA are acquired via fuzzy
logic interpretation of the relative values of the steady-state simulations of the corresponding rule-based
fuzzy cognitive map (RBFCM).
Findings – Through the use of RBFCMs, the following paper illustrates the possibility of incorporating
ethical components into machines, where latent semantic analysis (LSA) and RBFCMs can be used to model
dynamic and complex situations, and to provide abilities in acquiring causal knowledge.
Research limitations/implications – This approach is especially appropriate for data-poor and
uncertain situations common in ethics. Nonetheless, to ensure that a machine with an ethical component can
function autonomously in the world, research in articial intelligence will need to further investigate the
representation and determination of ethical principles, the incorporation of these ethical principles into a
system’s decision procedure, ethical decision-making with incomplete and uncertain knowledge, the
explanation for decisions made using ethical principles and the evaluation of systems that act based upon
ethical principles.
Practical implications – To date, the conducted research has contributed to a theoretical foundation for
machine ethics through exploration of the rationale and the feasibility of adding an ethical dimension to
machines. Further, the constructed AMA illustrates the possibility of utilizing an action-based ethical theory
that provides guidance in ethical decision-making according to the precepts of its respective duties. The use
of LSA illustrates their powerful capabilities in understanding text and their potential application as
information retrieval systems in AMAs. The use of cognitive maps provides an approach and a decision
procedure for resolving conicts between different duties.
Originality/value – This paper suggests that cognitive maps could be used in AMAs as tools for
meta-analysis, where comparisons regarding multiple ethical principles and duties can be examined and
considered. With cognitive mapping, complex and abstract variables that cannot easily be measured but are
important to decision-making can be modeled. This approach is especially appropriate for data-poor and
uncertain situations common in ethics.
Keywords Decision making and ethics, Natural language processing in machine ethics,
Rule based fuzzy cognitive maps
Paper type Research paper
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/1477-996X.htm
Rule based
fuzzy
cognitive
maps
231
Received 17 October 2015
Revised 12 April 2016
Accepted 14 April 2016
Journalof Information,
Communicationand Ethics in
Society
Vol.14 No. 3, 2016
pp.231-253
©Emerald Group Publishing Limited
1477-996X
DOI 10.1108/JICES-10-2015-0034
1. Introduction
The emerging eld of machine ethics endeavors to transform our increasingly pervasive
machines into safer articial agents. This recently emerging subeld of articial
intelligence (AI) is increasingly motivated by concerns regarding the dangers articial
intelligent agents may pose to humanity (McCarthy and Hayes, 1969;Anderson et al.,
2004;Yampolskiy, 2013;Bostrum, 2014), and focuses on the underlying design and
principles directed toward constraining lethal actions of autonomous agents, so that
their behaviors are bounded. Namely, machine ethics ultimately strives to develop
next-generation autonomous agents, capable of following ideal ethical principles in
decisions they make (Anderson and Anderson, 2007;Arkin, 2008), and it draws upon
interdisciplinary knowledge accumulated in both computer science and philosophy. By
focusing on the ethically acceptable behavior of articial agents, this new eld
distinguishes itself from the earlier work of computer ethics, a eld that has traditionally
focused on the ethical issues regarding the use of technology by humans (Anderson and
Anderson, 2007).
Within the machine ethics research community, it is commonly agreed-upon that any
articial moral agent (AMA) currently engineered would be an implicit ethical agent,
that is a machine capable of carrying out its intended purpose in a safe and responsible
manner as determined by its designer, and not necessarily able to extend its moral
reasoning to novel situations (Moor, 2006;Shulman et al., 2009;Chen et al., 2009).
Opinions within the eld, however, fragment on the desirability and feasibility of
developing an explicit ethical agent, that is an AMA analogous to an ethical human
decision-maker, capable of calculating the best action in ethical dilemmas, represent
ethical principles explicitly, operate effectively on its knowledge base and justify all
moral judgments and actions (Anderson and Anderson, 2007;Moor, 2006;Shulman
et al., 2009).
In this paper, two considerable challenges of machine ethics are explored. First is the
issue of the acquirement and incorporation of the necessary information needed to make
an ethical decision, which we attempt to solve by using latent semantic analysis (LSA);
second is the issue of the actual ethical reasoning and decision-making process, which
we attempt to solve through the use of rule-based cognitive maps (RBFCMs). The latter
problem is considered to be especially challenging due to the incomplete codication of
ethics (Anderson and Anderson, 2007). This incomplete standard has resulted in the
establishment of general disagreement on what moral structure AMAs should possess,
with diverging suggestions ranging from the application of evolutionary algorithms in
populations of articial agents to achieve “moral selection”, neural network models of
cognition and various hybrid approaches founded on ethical theories, such as virtue
ethics, Kant’s Categorical Imperative, utilitarianism, value systems inspired by the
Golden Rule and several others (Shulman et al., 2009).
The following study will demonstrate how RBFCMs can be used to represent
causality, where inputs and their effects are modeled using fuzzy operations (e.g. and, or,
if, then). The what-if scenarios commonly associated with cognitive maps are used to
mimic the thought processes associated with human thinking. This study will
demonstrate that the evolution of an RBFCM is iterative, where current values for each
of the concepts are computed with their inputs’ previous values. Here it will be
illustrated that the use of either crisp or fuzzy values for particular concepts is far
superior in capturing dynamic causal relations between concepts. The study will further
JICES
14,3
232
To continue reading
Request your trial