Good question. On the surface, it seems simple enough. Assign any standard you like as a demonstration of intelligence, and then ask whether you could (theoretically) set up an AI to perform it. Sure, it seems common sense that given sufficiently advanced technology you could set up a computer or a robot to do just about anything that you could define as being doable. But what does this prove? Have you proven the AI is really intelligent? Or have you merely shown that there exists a solution to your pre- determined puzzle? Hmmm. This is why AI futurist Max Tegmark emphasizes the difference between narrow (machine-like) and broad (human-like) intelligence.1
And so the question remains: Can the AI be intelligent, really, in the same broad way its creator is?
Why is this question so intractable? Because intelligence is not a monolithic property. It is not defined by any single measurement. Human intelligence operates on many levels, and displays infinite variety. Furthermore, it is revealed only within cultural milieu of understanding. Intelligence is discerned by humans in relationship, communicating and communing with one another.
The question of intelligence opens up a bouquet of interrelated questions:
Suppose that some future AGI systems (on-screen or robots) equaled human performance. Would they have real intelligence, real understanding, real creativity? Would they have selves, moral standing, free choice? Would they be conscious? And without consciousness, could they have any of those other properties?2
All these questions are really just different ways of trying to get at the same thing. Intelligence, understanding (meaning-making), creativity, morality, free will, and personal identity—these are all part and parcel of what it means to be human. In the Hebrew scriptures (Old Testament), the word nephesh implies all of these characteristics when it refers to humans. Nephesh is usually translated into English as soul, self, or person. It describes a living, embodied self with all of these properties present in whatever degree or potentiality the form of life presents itself.
The question of intelligence thus unfolds quickly into a set of interdependent questions about what it means to be conscious, to have a mind, moral agency, free will, and so on. These are deep questions that radiate from the heart of our beliefs about the nature of reality and what it means to be human. Your view of whether an AI could ever possess the properties associated with consciousness (and perhaps, even have a soul) thus depends upon your worldview.
A view through the lens of materialism
Daniel Dennett and Paul Churchland provide good examples of how these questions look through the lens of materialism. Churchland “denies the existence of immaterial thoughts and experiences.”3 Churchland calls his view “eliminative materialism.”4 Similarly, Dennett “denies the existence of ontologically distinct experiences, over and above bodily events.”5
These two philosophers base their arguments upon an axiomatic, a priori, and abject denial of anything that can’t be explained as strictly material. In other words, nothing exists unless they say so, because they get to define what is material and what is not. That means there is no such a thing as consciousness, as the word is normally used, as a description of the transcendent sense of self that most people seem to have. It might be fair to say of these philosophers that they believe their bodily functions are working just fine, but they don’t believe they are actually conscious beings. So be it. If that’s their mindset, then it seems clear they will be neither helpful guides nor open-minded discussants in any effort to understand any possibility that reality might ultimately transcend the a priori, fideistic constraints they have placed upon it by adopting the stance of “eliminative materialism.”
Rationales like those expressed by Churchland and Dennett go like this:
(A) (“Eliminative”) Materialism is the sum total of reality, by definition.
(B) Consciousness, as popularly understood, implies mindfulness (thoughts and experiences) that transcend materialism.
(C) Therefore, consciousness (at least as popularly understood) does not exist, because that would violate the axiom (A).
This is circular logic. We finish where we began, by denying the possibility of any reality other than that which can be explained by sheer materialism. Materialism is all there is, by definition. By this line of circular reasoning, there exists nothing which the materialistic philosophers cannot explain by dint of their materialism.
There have been attempts to salvage the metaphysics of materialism without giving up on the existence of transcendent realities. In order to do so, it is necessary to broach the boundaries drawn by strict “eliminative materialism.” We might label these attempts as “soft” materialism, to distinguish them from Churchland’s and Dennett’s “strong” materialism. Exemplars of this attempt are Warren Brown and Nancey Murphy, who have sought to “square the circle” and allow for the qualia of “soulishness” to emerge from within the framework of materialism.6 Murphy and Brown propose a metaphysics in which “soulishness” is an emergent property of the organism that arises from the material substrate of biological existence. They label their approach “non-reductive physicalism” (NRP). NRP thus attempts to explain why the concepts of self, mind, and morality are meaningful, even though they do not exist in any material, ontological way in and of themselves.
You are forgiven if this seems confusing. This sort of metaphysical tweaking of concepts obviously gets muddled quickly, because it’s not clear what axiomatic presumptions of worldview/faith have been invoked to sustain such rationales. People of different views on the meaning of these metaphysics often end up talking past one another, unable to mount any persuasive arguments. Margaret Boden sums up aptly, “The topic is a philosophical morass.”7
Religious believers and New Atheists alike probably agree on the messiness of the metaphysical morass. They are likely to agree also, regardless of their faith perspectives, that human life is meaningful and that morality is an appropriate concern. Nonetheless, the philosophical morass remains an obstacle to be navigated.
A path out of the morass?
The only way out of the morass is to recognize that truth claims do not stand on their own, aloof and cut off from the sea of meaning which grants epistemic access. In other words, truth presumes access to: (1) a way of knowing, and (2) a reason to trust our ability to know. Thus, the way out of the morass requires we recognize truth claims as speech-acts that take place within an epistemic frame of reference.
This inescapable connection between the field of knowledge and the act of apprehending truth calls for a posture of self-examination and testing of beliefs, to ensure coherence in thought and expression. This disciplined pursuit of knowledge, and scientific testing of theological truth claims is called, “dogmatics.” As Barth defines it, “dogmatics is the scientific self-examination of the Christian Church with respect to the content of its distinctive talk about God.”8
Whether the questions are addressed from within a religious, secular, or atheist perspective, the way out of the morass depends upon scientific dogmatics. The discipline of scientific dogmatics is a necessary ingredient for coherence within a process of self-examination of the affirmations which provide the ground of meaning for exploration and discovery.
Thus, the question of real intelligence needs to be preceded by a more interesting, more fundamental question: On what basis do you claim epistemic access to an understanding of what it means to be conscious, to have a mind, intelligence, freedom, and moral responsibility?
In the case of biblical faith, there exists a rational answer to that question, because personal and “scientific” truths are recognized as being contingent upon, and sustained within the truth of the self-revealing God. As T. F. Torrance explains in his scientific theology:
You know something only in accordance with its nature, and you develop your knowledge of it as you allow its nature to prescribe for you the mode of rationality appropriate to it. That is the kind of objectivity we adopt in all rational behavior whatsoever.9
To keep in step with this burden of rational inquiry into consciousness—to develop our knowledge of something in accordance with its nature—requires that we not treat consciousness as though it were a mere biological artifact of deterministic materialism in a godless universe. If consciousness is a qualia or experience of a human person, made in the image of God, then rational scientific inquiry will treat it as such. A biblical understanding of the created order gives scientists good reason to study the biological and material aspects of mindfulness, consciousness, and intelligence without blindly “defining away” the function of consciousness as a qualia of the soul.
So can an AI be intelligent, really?
Returning now to the opening question, “Is AI intelligent, really?” we see that some discussion of what we mean by terms like intelligence and consciousness needs to come first. It doesn’t make much sense to ask questions like this without considering the fundamental issue of epistemic access to knowledge and the sort of moral awareness that we expect of intelligent persons.
Since we rely upon this type of discussion with other people pretty much as a matter of course without even thinking about it, we probably need to place the same burden upon any intelligent machine before we start asking whether it is intelligent. It might make good sense to find out what the machine believes about the nature of reality before we pass judgment on its intelligence. Same goes for philosophers and theologians too.
1 Tegmark, M. 2017. Life 3.0: Being Human in the Age of Artificial Intelligence. New York: Alfred A. Knopf, 50-51.
2 Boden, M. 2016. AI: Its Nature and Future. First ed. Oxford, UK: Oxford University Press, 119.
3 Boden, p. 129.
4 Churchland, P. M. 1981. “Eliminative Materialism and the Propositional Attitudes,” Journal of Philosophy, 78: 67-90.
5 Boden, p. 129.
6 Brown, W. 2004. Neurobiological Embodiment of Spirituality and Soul, in Malcolm Jeeves, ed., From Cells to Souls – and Beyond: Changing Portraits of Human Nature. Grand Rapids, MI: Eerdmans. Murphy, N. 2006. Bodies and Souls, or Spirited Bodies? Cambridge University Press.
7 Boden, p. 122.
8 Barth, K., Bromiley, G. W., & Torrance, T. F. (2004). Church Dogmatics: The doctrine of the Word of God, Part 1 (Vol. 1, p. 3). London; New York: T&T Clark.
9 Torrance, T. F. 1971. God and Rationality. Oxford University Press, 52.
Follow this and additional works at: https://digitalcommons.spu.edu/works
Part of the Artificial Intelligence and Robotics Commons, Comparative Methodologies and Theories Commons, Epistemology Commons, Philosophy of Science Commons, and the Practical Theology Commons
Seattle Pacific University Digital Commons @ SPU