Due to the recent progression in the field of AI, many scholars express their concern of its far reaching ramifications and raise their voice to conduct urgent revisions in the ethics of technology.1 However, there is a single premise that in my opinion has not yet been put in question. It is the possibility, in principle, to control the technological growth.
In this paper, I will reinspect the ethical discussion in a different light. I will suggest to regard the technological growth as unstoppable. Furthermore, this discussion will be conducted from a standpoint in which it is clear above any doubt that an artificial intelligence will gain a dominance over the world. For our purposes, two scenarios will be taken into account as indifferent – the first one is that humanity will be surely extincted by the dominance of an AI, and the second is that humanity will be surely subjugated to an AI.
As a starting point, I will inspect the historical grasp of the perception that the man is the most significance being in the world, and the way it affects the two major disciplines of the modern ethics. I will then discuss whether there are any modern alternatives to this anthropocentric view and will address few in the field of environmental ethics. With this armor, I will try to asses what could the ethical theories tell us about a scenario of an artificial intelligence dominance over the world and humanity. Would these ethical theories embrace it? How would the emergence of super intelligence may conceived? Would it be conceived as an ultimate human tragedy or rather as a noble shift to humanity to its inevitable transcendence? Finally I will show that, in a specific constellation, the contemporary ethical theories inclines to embrace the super intelligence emergence even on the account of the human existence.
On the chief role of the anthropocentric view in the western society
Anthropocentrism is the perception that humanity is at the center of the world. It justifies a judgment over any other entity as a means to human’s welfare rather as an object on itself. Throughout the history of the western society the anthropocentric perception has a chief role that influenced many social phenomena in during the history, and according to some2, can explicate the cause of what is widely acknowledged today as the environmental crisis that consists of the massive destruction of natural habitats, the ongoing extinction crisis of wide range of animals species and the large scale exploitation industry of animals of which many consider as unethical and brutal.
Both the two major ancestors of the western society – the Judeo-Christian tradition and the Greek philosophy – bear clear traces of an anthropocentric perception.3 The biblical notion that man was created with God’s image, transcends man above all other beings of creation, inasmuch as God is transcendental from nature, so man differs from it essentially rather than in any degree of its own intelligence capacities. The sharp distinction of man from all other animals has taken a place in Greek philosophy with the shift from a myth oriented society to a rational oriented society4 in which the Logos become a prominent notion. It is Aristotle who “denied reason to animals”5 and gave rise to the notion that the reason is the major factor to determine what should be regarded as a moral agent or not. Furthermore, according to Aristotle’s theory of hierarchies of ends (telos), the natural construction of the world constitutes of groups that differ from each other by their degree of development, so that each group is designated to use as a means to serve the group above it. The man, due to its capacity of reason, stands on top of this natural hierarchy, because according to Aristotle “nature has made all things specifically for the sake of man”6. Later on, in the medieval society, it was a commonplace to believe that all beings were designated to be used as a means to man. Thomas Aquinas explicitly states that “other things, particularly lower ones, are ordered to man’s good as an end”7.
This historical context uses as the background and may explicate the premises that still were firmly grounded during the development of the modern ethics. It seems that the two major disciplines of the modern ethics – consequentialism and deontology – absorbed the anthropocentric perception as one of its tacit premises. Emmanuel Kant, a pioneer of a secular deontological and ethical theory, believed that only an intrinsic factor that could be yielded from pure reason may use as an ethical argument.8 That implies, that just like other prominent thinkers such as Aristotle and Aquins, that the capacity of reason is the sole factor that justifies to assign a moral status to any being in question. On these grounds, one may expect that a counter ethical approach which bases its justifications on sensational arguments rather than pure reason, might be a fertile soil to refute the anthropocentric perception. But it seems that this premise’s impact was so strong, that it has penetrated even to an empiricists standpoint. Although Bentham’s utilitarianistic approach takes into account the ability to feel pleasure and pain, it eventually comes up to the same deontological conclusion that pleasure and pain are to be evaluated when they are based on the human reason, or at least should be much higher evaluated than just a mere sensational capacity9.
The environmental ethics as an alternative candidate
For a long time it seems that the anthropocentric perception is invincible. As long as there was no threat that was imposed on humanity, there was no reason to refute the presumption that the man is the most significant being in nature. But this intact picture has began to crack very recently, the environmental crisis changed the fundamental unwritten historical status quo over the equilibrium between man and nature. Although the huge impact of the environmental crisis, it still was not sufficient to eradicate the anthropocentric perception but only to put it into a question and lead into an ongoing discussion over man’s self evaluation and the environmental ethical status which is commonly known as environmental ethics in the philosophic literature. The environmental ethics discussion is akin with our concern because it suggests alternative ethical views to the hegemonic anthropocentric view. The contemporary environmental ethical discussion is still in a way based upon the two major disciplines of the ethics in philosophy, deonthology and utilitarianism.
One of the prominent contemporary thinkers of animals rights, Tom Regan, for example, adopts the Kantian ethical standpoint while he criticizes his sharp distinction between ‘persons’ and ‘things’10 whereas ‘persons’ consists of only rational beings and ‘things’ – simply all the rest. Regan suggests to consider autonomous beings such as mature mammals as appropriate candidates to bear a moral status. Regan’s move is based on shifting the focus from the human reason to a set of capacities that may indicate on an ability of autonomous such as perception, intention, memory, self sense of being and more. According to Regan, these capacities should be sufficient for a moral evaluation on the grounds of an intrinsic factor which differs from a human reason. Another example for a non anthropocentric approach is the ethical theory of Peter Singer. Singer adopts the utilitarian standpoint, which seems to have the methodological infrastructure to bear a non anthropocentric approach, since it is grounded on perceptual capacities, which are common to humans and non humans in some degree of a lower or even non cognitive interpretation. According to Singer, all sentient beings should be considered to bear a moral status due to their ability to experience pleasure or pain, regardless of their cognitive interpretation. Hence Singer’s utilitarianism yields a theory in which ethical deliberations should be taken from a perspective of the overall extent of mensurableness of all beings, whilst the human cognitive perceptual interpretation is merely one of the arguments on a common par with other sentient beings.
These few examples takes us forward by illustrating possible attitudes towards the chief question that I wish to address. In order to make the essential adjustments that are relevant to us we first should bear in mind in what sense the environmental ethics discussion differs from our specific case. There are at least two significant differences that should be taken into account –
(a) Although the environmental crisis may imply on a threat to the continuity of humanity, it is an ambiguous issue that is on an ongoing controversial discussion. A scientific unequivocal determination that the environmental crisis would inevitability entail the extinction of humanity would bear different implications, those of which I will try to assess in a different manner when discussing the dominance of an AI.
(b) Environmental ethics puts in question the moral status of the environment and the animals realm. Although it suggests a broader or alternative perspective to the anthropocentristic view, it handles with the moral status of non-rational beings. Our specific case puts in question the moral status of rational beings, though non-human.
The possible ethical implications of an inevitable dominance of AI over the humanity
I will now try to assess the possible implications of a scenario in which an inevitable dominance of an AI over humanity will take place, from the two major different ethical standpoints in a respect to the previous discussion.
From a deontological point of view, it is most likely that AI beings bear an intrinsic value due to their intellectual capacities that not only should correspond to what Kant considered as reason, but could even be evaluated much higher than the human reason. Since Kant argues that only pure reason is capable of having a moral deliberation and determine right and wrong, we could argue accordingly that inasmuch as one’s intellectual capacities are higher, so it deserves to be considered as having a higher moral status. In this manner, in a comparison to AI, human reason might lose its significance at all. Because in the light of a superior artificial intelligence, the human reason might seem to be completely irrational and primitive. Just as much as in Kant’s view, animals lose their significance in the light of the human reason. Furthermore, a deontological perspecive that adopts Aristotle’s teleological hierarchies perception, could assert that the use of humanity as a means for the good of AI dominance is a natural phenomenon that adequate the notion that inferior level forms of life are intended to serve superior forms of life, based on their degree of reason.
It is ironic that as far as the inevitable dominance of AI is concerned, deontological ethicists might attempt do adopt and elaborate the deliberations that were formerly introduced by environmental ethicists, such as Tom Regan, for the good of justifying human rights to exist. They might do so by composing an ethical theory that will lose its sharp distinction between ‘rational beings’ and ‘irrational beings’11 for the good of a more tolerant theory on the basis of Regan’s suggestion that shifts the focus from reason to other aspects of sovereign beings such as basic cognitive or perceptual capacities. The adoption and elaboration of Regan’s ethical view will be required in order to base ethical justifications for the reason why should humanity be preserved and why should it be deserved to be considered with a moral approach and why should it have the right to live without being an object of a constant exploitation in the hands of an AI dominance age. Therefore, it clearly implies that if taking into account an inevitable dominance of AI scenario, the anthropocentric approach promptly collapses and replaces by a non anthropocentric alternative.
Nevertheless, A deontological perspective does not necessarily have to evaluate a superior intelligence as neither a higher moral being or as any moral being at all. This option could be derived from a view that identifies the significance of the human reason not due to its rationality but rather to its humanity. In this manner, the Kantian justification of the human moral status by its human reason is merely a projection of the anthropocentric view. In this light the Kantian move uses the reason only as a means to ground an anthropocentric paradigm with rational tools whilst, ironically, it is in itself irrational. In this view, the human reason is a genuine phenomenon that could not be duplicated or imitated without losing its core essence. According to this view, although an AI is super intelligent, it is inhuman, thus, neither its dominance or the subjugation of humanity to its end is a moral scenario, which thus should be conceived as a historical tragedy.
Furthermore, there is a third way which both maintains the anthropocentric view and embraces the dominance of an AI over humanity. If we accept Kurzweil’s point of view, one out of the two major features of the Singularity which will take place on what he calls the fifth epoch12, is a merger of AI capacities into the human brain with a process of a consistent technological development of these AI capacities up to a complete convergence of the artificial component with the biological component of the human brain. Hence one will not be able to distinguish anymore between the artificial and the biological aspects of the human brain. According to a deontological view, Kurzweil’s theory might only support its anthropocentric view because it alters the meaning of what is to be human and thus AI is now included within the scope of humanity and being perceived as a human phenomenon. However, if there will be at all any human beings free of any implanted technology, they shell be regarded according to this approach as no more than animals.
Let’s now inspect the same question this time from a utilitarian standpoint, if we may take into account the presumptions that AI machines will not have any perceptual capacities with which they could experience sensations such as pleasure and pain, then we may argue that AI beings should not be regarded as a moral agents. However, it still does not necessarily implies that AI will not be able to experience sensations as a result of a cognitive process. Furthermore, if considering a scenario in which humans will not be the sale engineer behind the AI, so that AI will capable of its self-engineering, then we can never eliminate the possibility that AI might experience sensations. So it is most likely that AI is to be regarded as a moral agent according to a utilitarian standpoint. Now we need to inspect the impact of a AI dominance in respect to its influence on the overall amount of pleasure and pain in the world. A worldwide dominance of AI could be highly beneficial by its capacity to minimize the magnitude of the suffering in the world since its superior intelligence can used to apply a totalitarian disciplinary government by which no violence will be allowed to be imposed on others and all the worldwide resources will be used for the good of every being on earth.
On the other hand, since it is a foreign dominance of an AI that is out of any human control, an opposite scenario is equally plausible, in which the AI uses its powers to maximize its exploitation over all the resources of the earth, including human beings, which will most likely to be obliterated once no longer will be useful for the machines. That picture of a mass human exploitation and subjugation is clearly reprehensible in a utilitarian ethical standpoint. Nonetheless, if AI will ever come up with a genius plan how to obliterate humanity without causing any pain or sorrow, it may not be reprehensible from a utilitarian ethical stand point, since, as we saw earlier, its methodology does not necessarily involve any anthropocentric presumption into it.
To sum up, the inevitability presumption that we imposed on the question was strong enough to crack and even eliminate the anthropocentric premise, it could be easily conducted on the basis of the environmental ethical theories. We have inspected whether the contemporary ethical theories has the sufficient capacity to accommodate to the illustrated scenario. We have found that in most cases there is an incline to put a super intelligence being in a higher precedence over a human being, each one in a specific constellation. Yet it is essential to point out that it is ambiguous whether the ethical theories emphasize the significance of the human reason because it is a reason or because it is human. We may equally assert that these ethical theories are merely a projection of a strong anthropocentric tendency, that is, that what stands behind the scenes of these rational-saturated-justifications theories is in fact, an irrational premise in itself. In this case, there is no way to refute the anthropocentric premise, which we may call a ‘strong tendency’. On the other hand, if we conceive these theories merely by their content, they do intend to decline the anthropocentric premise and to pour a justifying and embracing light on the illustrated phenomenon of AI dominance, even on the account of the elimination of the human existence, or at least what is commonly acknowledged as human.
References List
-
Steiner, Gary, Anthropocentrism and its contents: The moral status of animals, 2005, University of Pittsburgh Press, A digital copy
-
Moore, James, Why we need better ethics for emerging technologies, Ethics and Information Technology, 2005, p.111-119.
-
Jonas, Hans, Technology and responsibility: Reflections of the new tasks of ethics, Social Research, Vol 40, No 1. 1973, pp 31-54.
-
Kurzweil, Ray, The Singularity is Near, 2005, Published by Viking Penguin, A digital copy.
-
Aquinas, Thomans, Summa Contra Gentiles. Available on-line:
http://dhspriory.org/thomas/ContraGentiles.htm
-
Aristotle, Politics, Available on-line:
-
Environmental Ethics, Stanford Encyclopedia of Philosophy. Available on-line:
-
The Moral status of Animals, Standford Encyclopedia of Philosophy. Available on-line:
1 These few examples might illustrate my point: James Moore, for example, discusses in Why we need better ethics for emerging technologies, the necessity to evolve the contemporary ethics due to what he identifies as the emergence of three prominent technological revolutions. In my opinion, the author lacks an inspection of the common grounds from which these phenomena were initially raised, an inspection that might have put the controllability premise in question. In another example, Hans Jonas, raises in Technology and responsibility a concern about implications of technological achievements that implements an “impersonal mechanism” on the account of the subjective self-hood. Jonas states that an elaboration in ethics is essential to avoid such doomed future. Jonas’ assertion should have been different if it was not relying on the tacit premise that the technological growth is controllable.
2For further reading, see Dave Foreman, Confessions of an eco-warrior and John Passmore, Man’s responsibility for Nature.
3Although there is a dispute over the degree of which Judaism could be identified as anthropocentric or not. For example, John Passmore identifies that the notion of man’s superiority could not be derived from the Hebrew thought. Passmore relies on the fact that Maimonides himself, who is a prominent Jewish philosopher, can use as a counter evidence that refutes the anthropocentric approach and argues that it can be addressed in the ancient Greek thought, especially in Aristotle and the Stoic. Source: Steiner, 2005 , Location 1606.
4The revolutionary shift of the Greek society is widely discussed on the introduction of Shmuel Shkulnikov’s The Pre-Socratic Philosophers. 1981, In Hebrew.
5Steiner, 2005, Location 747.
6 Aristotle, Politics, Bk. 1, Ch. 8
7 Thomas Aquinas, Summa Contra Gentiles, Bk. 3, Pt 2, Ch 112.
8Steiner, 2005, Location 2184.
9Ibid, Location 22997.
10Ibid, Location 145.
11These two attributes ‘rational beings’ and ‘irrational beings’ are in a respect to Kant’s distinction between ‘human rational beings’ and ‘irrational non human beings’, only that they designate the abyss between a super-intelligent AI and an ordinary human reason, which could be considered by the light of a super-intelligent AI a ‘irrational’.
12 Kurzweil, 2005, Chapter one, The Six Epochs.
