Abstract
Virtual reality (VR) has both a technical and a cultural dimension. Culturally, it is can be defined as a simulation in which we are invited, or perhaps persuaded, to amend our belief in what is real. VR is significant because it follows artificial intelligence and corresponds to the latest step in the gathering crisis of modern philosophy. There are therefore many problems in developing an ethics of VR, but it is possible to adopt pragmatic solutions. For example: it is possible to appeal to past ethical codes, it is possible to propose a philosophy of the artificial and it is possible to re-analyse our present dilemma and assert a humanist philosophy. This latter approach is adopted and some comments on the form and content of a possible ethics of virtual reality are provided.
Introduction
This paper is an attempt to examine some of the ethical issues that arise from the potential use of virtual reality (VR) systems, to analyse why it is so difficult to derive a clear ethical stance with respect to these issues and, finally, to seek some ethical principles despite these difficulties. In order that the discussion take place in some context, let us first consider some possible areas of concern with respect to the future use of VR systems.
* VR is being developed for new devices in video parlours. What is the likely effect of spending an extended period in a virtual environment where it is possible to completely discount the worth of other people and to deal with them only by means of militaristic or sexual aggression?
* Personality testing VR systems could be developed and used as a selection device. For example, in place of the usual job interview, applicants would be made to enter a virtual testing scenario and have their reactions recorded. Decisions about their future would be taken on the basis of these recordings.
* VR could be developed as a punishment for offenders. Instead of spending time in prison, an offender could be sentenced to a period in an "appropriate" virtual reality.
* Aspects of a real person's life could be recreated as a virtual world for other people to "experience", raising important questions about personal privacy.
* VR could be developed as a form of psycho-therapy. There could be virtual experiences which aim to change people's attitudes, and ones that simply aim to tranquillise people.
* VR systems, like any other technology, can malfunction and have serious effects upon the user. Our first problem is to define what exactly constitutes a malfunction in a virtual world. Our second is to decide what effects are attributable to the malfunction. Our third is to determine who would be responsible for any bad effects.
* There are various experiences, such as flight simulation and lengthy periods in submarines, that can temporarily disorient the user after the experience has ended. Similar effects are probably associated with leaving a virtual world. There is likely to be a particularly dangerous period when users adjust to the real world after their VR experience and when they have very little control over their actions.
* VR systems are also being developed for the training of operators and the testing of products. Insofar as the model within the VR system cannot perfectly reconstruct the real world scenario, there is a danger that people trained in this environment or products developed in this environment may fail when confronted with real situations.
These are just some of the questions that will need to be addressed if we are to adopt a responsible approach to VR systems, but addressing them is not likely to be easy. We will start by trying to clarify what we mean by "virtual reality" and what the preconditions are for any ethical intervention.
What is virtual reality?
There are two aspects of virtual reality (VR): it is both a set of technical projects and a utopian concept. From the technical perspective VR is very small, there are perhaps twenty projects that have any claim to be serious VR (see, for example, Eisner 1991) though it could be claimed that none of these achieve the required level of belief. Despite these few examples, the cultural impact of VR is enormous. There are thousands of people who are excited by the concept of entering "other worlds" (Haggerty, 1991).
In this respect, VR is very similar to artificial intelligence (AI). It too has had limited technical success, but there has been enormous interest in AI as a concept. The major difference between these two concepts is that whilst AI was primarily concerned with knowledge, VR is concerned with reality. (In philosophical terminology, AI is concerned with epistemology whereas VR is concerned with metaphysics.)
It took several decades of work in AI before ethical questions were raised (Weizenbaum, 1976; Floyd, 1985; Torrance, 1986; Whitby, 1988 ). If we want to raise similar questions with respect to VR then we must first provide an adequate analysis of the concept of VR. This analysis will concentrate mainly upon the cultural concept of VR (i.e. the possibility of visiting "other worlds", or "other realities") rather than upon technical advances. This is because ethics must be aimed primarily at affecting the actions of people who are pursuing a social goal. The analysis of this goal is therefore our first aim.
Superficially, the concept of "virtual reality" contains a contradiction. The
word "virtual" implies at least an element of unreality and so we when we talk
of VR we seem to be referring to an "unreal reality". Further analysis,
however, suggests that in VR the word "virtual" is used in a slightly different
sense:
"That is so in essence or effect, although not formally or actually; admitting of being called by the name so far as the effect or result is concerned." (Oxford English Dictionary)
Virtual reality therefore has the effect of reality upon us, though we recognise that it is not properly real. VR is best described as a simulation, as opposed to a representation (Baudrillard, 1983). If I make a model of something, say a chair, then my model is a representation if I never lose my belief that it is the original chair that is the real object. On the other hand, if I make a model of something, say the surface of the moon, and in navigating the imaginary terrain I come to believe that the model is real, then I am in a simulation. The term "virtual reality" is a precise expression of this latter concept for we are, at the same time, admitting the fact that we have created a model (which is unreal) and admitting that we are treating it as reality.
The significant point about this definition is that VR is essentially subjective. VR is an experience and not a piece of technology. I can curl up with a good novel and claim to be in virtual reality, whilst I can don the most expensive headset and data glove yet remain perfectly aware of the fakeness of what I observe. This is important because it follows from this definition that many symbolic structures in society can be viewed as "virtual reality" (for example, cases where a computer model of an organisation does not reflect the underlying reality). We must therefore understand virtual reality as being the expression of a deep philosophical problem caused by our commitment to symbolic structures whose existence now obscures the reality of the underlying object.
The role of ethics
Given this concept of "virtual reality" we must next analyse the role of ethics in regulating behaviour. Traditionally, the idea of free action has been important for ethics. On the one hand the restrictions placed by the Law are seen as legal obligations which restrict this freedom and therefore place one limit on the scope of ethics. On the other hand, individuals must be free to make their own decisions wherever possible, so not everything that is disapproved of can be considered unethical. For some moral philosophers, it is possible to advance ethical principles which aim to influence human behaviour within a social context but which neither overlap the Law, not interfere unnecessarily with individual freedom. This can be done by accepting that people generally act in order to achieve personal happiness, and admitting a social dimension in which people also wish to be seen by others as deserving that happiness (Acton, 1970). Another way of expressing this would be to say that, as well as wishing to achieve happiness, people wish to be seen to be acting legitimately (as opposed to just legally). In this sense, an ethical dimension legitimises an action regardless of whether the action succeeds in achieving the state of happiness sought.
Apart from such motives, there may be more immediate benefits of regulated action. New technology inevitably breeds the spirit of the frontier, the unbridled push into new territory, but it does not take long for practitioners to desire some elements of regulation. The need for technical standardisation is one aspect of this, but there are often strong pressures to regulate behaviour as well. There may be a variety of reasons why a particular group might seek a code which specifies what is considered to be appropriate action (Holvast, 1992). Firstly, such a code will help a practitioner act with some confidence in a situation in which she or he is uncertain about what is correct. It can advise that in a certain type of situation, then do this. Secondly, it will help limit the responsibility of a practitioner and may help establish innocence when the behaviour of the practitioner has been challenged. Thirdly, a code may help eliminate unfair competition by establishing a minimum standard below which it is not permissible to fall. In short, a code of ethics (or code of conduct) will help determine which actions are deemed to be legitimate and which are not and it will try to eliminate lengthy and inconclusive argumentation whenever an action is challenged.
Traditionally, we think of philosophy as consisting of a group of sub-disciplines - philosophy of religion, aesthetics, ethics and epistemology - centred around metaphysics. It is often forgotten the extent to which these sub-disciplines are inter-related. Sloman (1985), for example, relates ethics to epistemology when he says that the question of whether a machine can be conscious is essentially an ethical question as well as a factual one. Vaux (1986) relates metaphysics and epistemology when she points out that we do not logically deduce that other people exist, their existence is simply a fact that we experience. There are many other relationships we could observe, the clearest perhaps being that without a belief in God there is not much future in a philosophy of religion.
Ethics can therefore only be clear in the context of a specific set of beliefs about what exists and how we can obtain knowledge. It is my thesis that VR is a reflection of a deep crisis in our concepts of knowledge and existence and so we cannot expect there to be simple clear cut answers to ethical questions. This does not mean, however, that we are powerless to formulate ethical principles, only that we cannot expect them to be universally accepted.
VR & the crisis in philosophy
The idea that modern philosophy is in crisis is not new. Some postmodernists express this by saying that we are at the end of the project that began with the Enlightenment (Dews, 1989). Laufer (1991) has shown how philosophy has moved through three stages since the Enlightenment: the first (from 1790 until 1890) was dominated by Newtonian science and Kantian philosophy; the second (from 1890 until 1945) was dominated by Comtean positivism and what we would call "modernism"; and the third (the period since 1945) is the period of deepening philosophical confusion and the emerging concept of the "artificial".
If this analysis is correct, then the emergence of virtual reality at this point in time is a reflection, not just of technical, economic and political developments (which are of course also very important) but of the fact that our traditional philosophical system is now collapsing at its most central point - metaphysics. Our concern with the ethics of virtual reality is therefore doubly difficult. Ethics has been severely attacked and has been in a state of confusion for at least fifty years (Ayer, 1936), and virtual reality is a reflection of deep philosophical confusion.
How to act in a crisis
Although we are in the midst of a large philosophical crisis, life goes on and actions have to be performed even if we do not have the philosophical certainty to know how to act. Furthermore, if our actions involve other people then we have to convince those people of the legitimacy of our action, i.e. give a symbolic explanation that is acceptable. This latter task becomes very difficult in the age of philosophical crisis. The legitimisation of actions tends to become a lengthy business.
Without a consensus we must resort to pragmatism, that is to say we act according to our own interests and then seek to legitimise them. The question becomes, how can we legitimise our actions? In general we have three possible ways: we can appeal to an old system, we can try to anticipate a new system, or we can look for some possible alternative interpretation of the present.
Referring to the old
When faced with a breakdown of the system of legitimacy we can attempt to legitimise our actions by appeal to the philosophy of some previous age. This is especially attractive to those in prominent public positions who may have to constantly legitimise their actions to large numbers of people.
For example, the period of 1790-1890 corresponds to the philosophy of the free market in which almost every action is considered legitimate and the market is left to decide whether it is the correct action. The major ethical principle of this period was Kant's Categorical Imperative. This concentrated on the actor's intention, and stated:
"Act only on that maxim through which you can at the same time will that it should become a universal law." (Kant, 1948)
Given the commercial content of many of Kant's examples, one can interpret this maxim in the spirit of free-trade as prohibiting only those actions which might give unfair advantage (such as not honouring promises). On this basis, any attempt to develop virtual reality is prima facie legitimate and the market will be the main arbiter of its fate. It is consistent for present day free-marketeers to adopt an ethical code that states that ethics is either irrelevant or should seek only to eliminate unfair competition.
There are also those who would return to the age of positivism with the strict separation of ends from means. Within this philosophy each discipline looked after its own methods and the ends were justified by reference to either God or Progress. The ethical standard within each discipline ensured that its practitioners were properly qualified in the methods of the discipline, carried them out diligently and did not undertake work for which they were not qualified. It is consistent for present day positivists (e.g. engineers) to adopt an ethical code that states that ethics should seek only to guarantee the competence of practitioners.
Adopting the new
Another way of attempting to create legitimacy is to refer to the formation of a new philosophical framework which, it is claimed, is about to replace the one that is in crisis. There are several contending versions as to what is emerging in the 1990s, the front runners being "postmodernism", "the artificial" and "systems science". As each have so many variations, the labelling of particular views is often unreliable as to content.
The concept of "the artificial" has emerged as the result of two related processes: the first is the fragmentation of single philosophical concepts into several competing terms, and the second is the reification of those concepts. The clearest case has been that of knowledge which was, in Kantian philosophy, a single unified concept. Now we have not only opinions which are given some epistemological status (in modal logic), but also information and data. The practical side of knowledge is given another name, intelligence, which has been separated from its human subject, so that we can now talk about intelligences as entities which can, in theory, co-exist with any type of object. These twin processes of fragmentation and reification were the main cultural developments that made the conceptual programme of artificial intelligence so important.
Virtual reality entails a repeat of the same processes with respect to existence. First we observe the fragmentation of the notion of existence, which began with the study of modal logic (Hintikka, 1962). Modal logic introduced the notion of a "possible world", being a discrete state of affairs that might have been. One writer has even proposed the notion of "partial possible worlds" ( Rappaport, 1987, p.67). It is to be noted that the concept of "possible world" has not gone unchallenged (see Nagel (1979, p.171) who argues for the "existence of facts beyond the reach of human concepts").
Further evidence of the fragmentation of our notion of existence is presented in articles such as Hirst (1989) which suggests that the verb `to be' can be used to express at least nine different forms of existence. Following this fragmentation we expect a move towards reification, that is the separation of "realities" from human experience. The concept of virtual reality is the embodiment of this development.
The ethical dimensions of these developments are as yet unexplored, but there is much to be concerned about. If ethics requires an area of human activity between what is required by law and what is considered legitimate freedom, and if it also requires a social context for our action, then we must ask whether the virtual worlds that are proposed include these necessary preconditions for ethical judgment. Of course, there is always the option for the philosophy of the artificial to redefine ethics, but I have yet to see any serious effort in this regard.
A further problem is that ethics, as presently conceived, requires the notion of individual responsibility which, in the age of the artificial, is also in crisis. Individual responsibility has traditionally been based either on the intention of the actor or on the outcome of the event. With complex systems that give rise to "virtual realities" then we need to know whether the same bases apply and, if so, in which world we are to judge the intention of the actor or calculate the outcome of the event.
Systems science has given us one important idea with the potential of an ethical dimension, and that is sustainability. Countering the modernist preoccupation with progress, Wiener (1961) proposed the goal of building sustainable systems. Rather than seeing human activity as slowly contributing towards the construction of more and more universal knowledge, Wiener held that all systems (knowledge included) naturally tend towards chaos and disorder. Intellectual work, for Wiener, entailed constantly trying to reconstruct the symbolic structure which we use to understand the world. The idea of elevating sustainability to an ethical principle is interesting, though it could be argued that because sustainability is not strictly a matter of choice it is not strictly an ethical notion. On the contrary, it seems to me that it is precisely the sort of ethical maxim that Kant would have approved of.
Humanism & VR
A third approach to establishing legitimacy is to identify the cause of the philosophical crisis and to try to resolve it. If, as Laufer (1991) maintains, the crisis is due to complete confusion of nature and culture then this requires a direct challenge to the philosophical position that permits the cultural concepts of artificial intelligence and virtual reality. Central to that philosophy is Turing's paper, Computing Machinery and Intelligence which proposes a machine/being that obeys both the laws of nature and the laws of society. The essence of the response must be based on the fundamental incomparability of human and machine behaviour. That is, a new form of humanism must evolve based on the fundamental standpoint advocated by Norbert Wiener when he argued that it was morally wrong to see human labour as in any way in competition with machine labour (Weiner, 1967). It is unfortunate that Wiener's humanitarian views were increasingly marginalised during the Cold War period in favour of von Neumann's positivist approach (Heims, 1980).
Weizenbaum adopts a similar humanitarian stance to Wiener, objecting to certain computer applications because they represent "an attack on life itself" (Weizenbaum, 1976 p269). He goes on to claim that it is immoral to use a computer system to replace "a human function that involves interpersonal respect, understanding, and love" (ibid).
Most writers on the ethics of computing have placed themselves in this tradition. Whitby, for example, warns against the idea that computer technology raises new ethical principles,
"the most consistent and morally most desirable way of dealing with the problems associated with robot morality is to regard it as a set of moral problems which are continuous with the areas of human and animal morality." (Whitby , p.177)
Christiane Floyd has synthesised several ethical positions into a few clear principles (Floyd, 1985). The first, which she attributes to Weizenbaum, is the principle that one should not do things with computers that one could not accept responsibility for doing without computers. Like many principles it is easy to foresee difficulties in application. For example, does the principle have any force when the task could not be achieved without the use of computers? Such quibbles should be discounted for the central meaning of this principle is that the computer should not be used to create exceptions to general ethical principles. This principle could and should be applied to VR.
Secondly, Floyd states that the use of computers is only defensible if,
"there remains sufficient scope outside the computer application for other human faculties and forms of experience not to degenerate." (Floyd, 1985 p.3)
A recent editorial in the medical journal The Lancet raises the same issues specifically within the context of VR:
"Continuous exposure to VR will impoverish those aspects of life that determine social development, interpersonal insight, and emotional judgment." (Lancet, 1991, p.283)
Underlying all these comments there is a belief in the primacy of human life and human experience, and especially of inter-human experience, a point also raised by Ardigo (1989). This view is not an argument against the use of all technology, for technology can be used to increase inter-personal communication (e.g. the telephone) and even enhance it for some people (e.g. the less physically abled). It does, however, bring into question the desirability of creating "alternative worlds" in which the possibility of a human social perspective is excluded.
In addition to these principles we need a specific ethics of VR which addresses the essential nature of VR itself. The nearest to this has been Floyd, who proposed a third principle: that the use of computers is only defensible if
" computers are used in applications where computation and symbol manipulation are adequate ways of dealing with reality." (Floyd, ibid)
Rather than leading us out of our present dilemma, this formulation begs the whole question of VR (though, to be fair, Floyd wrote this several years before VR was conceived). In addition to the responsibilities we have as developers and implementors of technology there is a central ethical question of VR which is, "What is the nature of the responsibilities one has when offering a new version of reality?" To address this question properly we must step outside the ethics of technology and refer to other ethical traditions.
The first is the responsibilities of any creative person who is presenting a film or play or novel. In general, we believe that freedom of speech is important, but it is possible to retain this freedom while still arguing that one should follow certain moral principles. The freedom to promote sexism, racism or religious blasphemy in creative works have all been called into question recently. There is no reason why VR should be exempt from any ethical principles which refer to creative works in general, though the matter of what those ethical principles should be remains a matter of dispute.
The second is the field of medical ethics. In my proposed definition of VR I included the phrase, "we are invited, or perhaps persuaded, to amend our belief in what is real". The distinction is a real one, for the responsibilities of an author whose work leads to certain emotions in the reader are less than those of a person who administers a drug knowing that the same effect will be induced. The responsibility is less because the subject who experiences the emotions has far greater control over their entry into the virtual world in the first case than they have in the second.
As the technology of VR approaches the point where it can ever deliver an experience which is immediately convincing then the responsibilities of the author of that reality are no less than the person who administers a consciousness controlling drug, and the ethical principles they work under should be no less severe.
Bibliography
Acton H B (1970) Kant's Moral Philosophy. Macmillan, London
Ardigo A (1989) New technology and social citizenship. Paper presented to the Intl Workshop on Human Centred Sytems, Brighton Polytechnic, 22-24 Sept 1989
Ayer A J (1936) Language, Truth and Logic. Victor Gollancz, London
Baudrillard J (1983) Simulations. Semiotext(e), New York.
Dews P (1989) From post-structuralism to postmodernity: Habermas's counter-perspective. In: L Appignanesi (Ed) Postmodernism: ICA documents. Free Association Books, London, pp 27-40
Eisner R (1991) Researchers see a wealth of applications for virtual reality. Scientist, 5, pp 14-16
Floyd, C (1985) The responsible use of computers: where do we draw the line? CPSR, Palo Alto
Haggerty,M. (1991) Virtual Reality dominates SigGraph. IEEE Computer Graphics and Applications. 11 pp 14-15
Heims S (1980) John von Neumann and Norbert Wiener: from mathematics to the technologies of life and death. MIT Press, Cambridge, Mass
Hintikka J (1962) Knowledge and Belief. Cornell University Press, Ithaca, New York
Hirst G (1989) Ontological assumptions in knowledge representation. Proc 1st Intl Conf on Principles of Knowledge Representation and Reasoning. Morgan Kaufmann
Holvast J (1992) Code of Ethics. Presentation to IFIP WG 9.2. Namur, 11 January 1992 (unpublished)
Kant, Immanuel (1948) The moral law, Kant's groundwork of the Metaphysic of Morals. Hutchinson University Press, London
The Lancet (1991) Being and believing: ethics of virtual reality. The Lancet, 338, pp.283-84
Laufer, Romain (1991) The history of computers: an epistemological point of view. In: J Berleur, A Clements, R Sizer & D Whitehouse (eds) The Information Society: evolving landscapes . Captus/Springer Verlag, Toronto/New York
Nagel T (1979) What is it like to be a bat? In T Nagel, Mortal Questions. Cambridge University Press, Cambridge
Rappaport W (1987) Belief Systems. In S Shaprio (ed) Encyclopedia of artificial intelligence. Wiley, New York
Sloman, A (1985) Proceedings of the 9th Intl Joint Conf on A I.
Torrance S (1986) Ethics, Mind and Artifice. In K Gill (ed) Artificial Intelligence for Society. Wiley, Chichester, pp 55-72
Turing, Alan (1950) Computing Machinery and Intelligence. Mind LIX, No 236
Vaux J (1986) AI and philosophy: recreating naive epistemology. In K Gill (ed) Artificial Intelligence for Society. Wiley, Chichester, pp 73-78
Weizenbaum J (1976) Computer power and human reason: from calculation to judgment. Freeman, San Francisco
Whitby, B. (1988) Artificial Intelligence: a handbook of professionalism. Ellis Horwood, Chichester
Wiener, N. (1961) Cybernetics: or control and communication in the animal and the machine. MIT Press, Cambridge, Mass
Wiener, N. (1967) The human use of human beings. Avon Books, New York