Read the essay of AUBG student Elizabet Ivanova who won second prize in this year’s University Council Competition. The topic for 2023 was “Is there a human attribute that AI can never surpass?”.
One might say that the answer to this question seems obvious enough – emotions, of course! But is it really that simple? Emotion AI, or affective computing, which sounds like a plotline in a dystopian novel, is progressively improving each year and might be well on its way to acting out emotions better than humans can.1 Moreover, to call emotions a “human” attribute is to neglect billions of other beings whose emotional states we are yet to understand and learn to appreciate.2 So no, I would not narrow down the answer to emotions. Instead, I would focus more on the general concept of the psyche of which current scientific research or any religious or esoteric knowledge still, to this day, knows very little. Anyone who has used a computer in their life would know that it will not do something if one simply stares at it, but if instructed, the computer performs its actions well (most of the time). So how would one instruct a computer to own a soul if we ourselves do not have the slightest idea of what that is and how it came to exist?
I do not refute any of the currently available and trendy, to some extent, advantages of artificial intelligence – in fact, I admire the people with native wits who are capable of producing such complicated mechanisms and algorithms.AI has never been better, and this is evidenced not only by its surging popularity and increasing universality but also by adoption statistics.3 The most recent and perhaps well-known at this point example of how widespread and accessible AI has become is ChatGPT – the new, more refined Google. Instead of billions of results and a second page that almost no users click on (let alone the ten completely irrelevant pages after that), ChatGPT converses with you. It tries to give you the most accurate information currently available and tailors it to the question as relevantly as possible. That alone makes it a more “human” experience and provokes more and more people’s interest to try it out and use it even on a daily basis; which in turn, improves the AI’s performance and “humanness” since it learns from patterns and newly acquired data from users in conversations. But is being human-like enough to replace an actual human companion in our professional or personal lives?
I would argue that the human resemblance of AI is rather uncanny and rather terrifies than predisposes one to perceive the machine close to our own nature. There are a number of studies describing and attempting to explain the “uncanny valley” – that unsettling feeling humans experience when confronted with not human-like enough robots.4This feeling appears to me to be the normal response of a creature that fears its nearing replacement as the “most evolved” or “intellectually superior” species on the planet. But that is only because we view ourselves as just that – powerful and superior. What lies beneath the superficial intellectual supremacy is the true supremacy of the infinite and undefinable sentience experienced through the psyche. That kind of supremacy is the one attribute that a machine cannot replicate. The machine learns from what it has at its disposal; even newly generated content is still nothing but a compilation of memorized resources that have already been produced by humans. AI learns from data we have given to it; all the text ChatGPT generates for your history class paper is merely a collection of other students’ papers stored in various databases and forums online. Whether that is plagiarism is another issue; what matters here is that this content is not original, it is acquired. To be clear, humans create. Machines simulate. Mimic. Compile. Unless instructed, a machine would not go ahead and invent new art forms or schools of thought. Humans would, even if we never had an initial understanding of those things.
Creation is at the essence of what it means to be human, quite literally.5 The word “human” is derived from the Latin “humus” which means ground or earth. And that is precisely where everything is born and dies. The earth is what brings and maintains life. I would like to emphasize that reproducing life in the purely physical sense is not the main point here. It is the life sparkle that we transfer once a new human is created – that particle of transcendentality which we have not been able to explain, document, or define for the thousands of years we have spent on this earth. Many attempts have been made in the forms of religions and conspiracies to shape out the notion of life (i.e., the soul), yet nothing conclusional has been achieved, nothing dogmatic, nothing close to an undeniable, absolute, and single truth, just like the fact that the Earth is round (even if certain specimens question even this basic truth occasionally). Scientifically, there is no soul – there are just neurons, brain chemicals, and complicated biological processes. The soul is nothing but a rejectable hypothesis.6 As a person who has a strong affinity towards science, I have always found this lack of research effort in the psyche field troubling but also understandable. Science is rather reluctant to say it does not know something. But whatever it has established as proven facts, it tends to develop well and achieves impressive progress. Like in the case of AI.
But that is precisely the problem. And the answer to the question this essay is attempting to answer. The available scientific knowledge is just enough to produce more of what has already been recognized as truth – data. Information, provable hypotheses, structured algorithms, and systematic approaches are the driving force of this unbelievable human progress, modernization, digitalization, and all the other -zations. And machines have long been on their way to facilitate that progress and become a substantial, if not prevailing part of it.7 Why? Because they deal with known datasets; because their capabilities have scope; because even in the sea of data, there are limits. What does not have limits cannot be efficiently replicated because where does it end or start? Machines might mimic and replicate to perfection, but they cannot generate non-data. What would non-data mean to a computer? Perhaps gibberish code which they would not understand and consequently, terminate the program or just ignore and continue to the next loop iteration. In its unsuccessful attempts to define, document, and limit the notion of life, science does the same. It proceeds without that notion and discards it as a mere hypothesis.
Humans like order and structure – in states, cities, communities, families, relationships, and tasks, but when the very foundation of that structure is unclear (life itself), threatening to transform entirely everything we have ever known, it is only natural to try to avoid that foundation. Perhaps this aspiration for structure also influenced our desire to build machines – to structure structures that can better structure our current structure. One does not need to possess expertise in psychology to understand the fear of the unknown and inexplicable, and how it shatters one’s security, exposes vulnerability, and obliterates the sense of structure. That is why we rely so intensely on data; it gives us a sense of structure we would perhaps never experience in our spiritual life. But since our psyche is not data, we cannot feed it into the AI’s dataset. And even if it was some sort of data, we could never comprehend its parameters, structure, or any kind of determinators. So, if we cannot scope the entirety of the psyche data, how could we expect the computer to perform any kind of operations with it, let alone surpass it?
More importantly, the whole initial purpose of the computer was to automate calculations.8 What is there more soulless than performing mathematical calculations? As much as mathematics has contributed to human development, it remains a scientific field of strict rules and limited free thinking. The inexplicable cannot be converted into afunction or if it can be, it would likely be too simplistic. And since the solid and stringent world of mathematics and logic are at the core of computing and therefore, of AI, AI is very unlikely (if not impossible) to escape the boundaries of cold calculations and transcend into a being.
Of all the human inventions, language is my favorite because it attempts to grasp all things and experiences, no matter how nonsensical or illogical they are. An example of this grasping is exactly the word being, which means not only existence but also “the nature or essence of a person.”9 Without dwelling much into the philosophical meaning of what the essence of a person entails, we know that “being” is the gerund form of “be,” which is to exist, to happen. We, as humans, happen, we go through a life cycle of living and simply being; we are the representation of life which happens only because we are making it happen. It could be argued that the concept of life itself is our doing. We create and are existence. Our minds, as rational builders, are the engines of what our souls, as irrational constructors, are striving for – being. I had planned to refrain from quoting Shakespeare’s widely popular Hamlet phrase due to its overuse and frequent misinterpretation, however, that indeed is the question. To ask ourselves whether we are or not is a quality only beings can possess for this being is our essence. And even if we do not completely understand what that being encompasses, we can feel it. What can an AI feel? Are its mechanical components hurting?
Do its classes and functions contradict each other in a heated existential discussion? How would an AI have any concept of being if its software and even hardware have never lived nor happened?
Another semantic peculiarity to mention is the word animation. The (sur)realism of the uncanny valley discussed earlier could be masked by the ever-evolving animation techniques that “breathe life” into all kinds of technologies, including AI, and including the ability of AI itself to further improve those animation techniques. “Animation” comes from the Latin word “anima” meaning soul or spirit. Even with the scientific understatement of the presence and significance of the soul, animating different types of objects, entertainment outlets, robots, etc., ironically means equipping those with a soul. Why are human-like machines unsettling to look at and interact with? Because a number of them attempt to appear alive which by itself incorporates the necessity for a soul. Seeing these inceptive attempts of piles of metal parts to artificialize a soul would likely sound preternatural to any naturally made 100% organic soul.
But of course, lacking a soul is beneficial in many ways and the inability to exhibit this human attribute does contribute to human progress. For instance, the human tendency to get attached to objects and other beings may prove to be detrimental to our health and wholeness upon separation from the object of attachment. This weakness is something a machine does not experience; the computer retains its rationality and full self-control because the concept of attachment is likely as foreign to it as is the concept of the soul. Furthermore, AI’s spectacular capacity to store and process mind-boggling amounts of data is truly something it outshines humans in and proves to be of great use. Some of the biggest accomplishments are in the medical field, such as advances in medical imaging that significantly improve the accuracy of MRI scans.10 Other advances in the field involve the early detection of various diseases such as Alzheimer, cancer, diabetes, and chronic heart disease.11 Yet other valuable achievements include the surgical assistance of AI which is becoming more and more prominent and would likely be a part of common practice in the near future.12 And since AI cannot be disturbed by empathy and other feelings that could overwhelm it and consequently, disrupt the process of research or treatment, AI is practically a perfect fit for science and scientific progress.
Speaking of perfect fits, AI’s presence and continuous improvement do not imply it needs to surpass anything human. On the contrary, AI can serve well as an ally to humanity. Referring back to the concept of anima, besides the word being the origin of “animation”, the Anima is also a famous concept in Jungian psychology. Together with the Animus, these two notions form a dual archetype outlining the two sides of the human psyche and/or personality.13 While both the Animus and Anima are psyche components, they differ greatly. The Animus is the representation of intellectual supremacy and rational decisions, while the Anima is the fragile, highly sensitive part of the irrational side of the soul. Despite the Animus/Anima archetype being primarily applied to psychology, this dual concept can also be paralleled to the humanity/AI dichotomy. Humans’ imperfections like emotional dependence and impulsive actions can be curbed byAI’s faultless performance. Humanity’s aspiration to create novel things, and invent thoughts and devices never heard of before out of the impulse of our irrationality can be facilitated by the collected calculations of a carefully constructed AI. There is no competition that juxtaposes humanity’s irrationality against AI’s rationality. Or at least, since we are asking such questions like “Is there a human attribute that AI can never surpass?”, there is such competition, and notably, we have created it ourselves.
This same juxtaposition is one of the many things AI is not capable of. Because constructing this same juxtaposition involves not as much rational thinking and computations as it does fear of replacement and fear of artificializing life. We do not need to be religious to feel disconcerted when a metal construction, with a microchip instead of a heart, attempts to conceive even the most basic form of life in the act of conversing with us. We do not need to be scientists to establish that this uneasy feeling is not quite right and is in evident contradiction with our essence as beings.
The psyche is the one thing a machine can never replicate, and thus surpass because there is nothing to replicate; the psyche is an entire eternity contained in a single living being – be it homo sapiens or other animals. This eternity’s limits or lack thereof cannot be reproduced even by an infinite number of calculations. Describing and defining the wonder of living, existing, being, experiencing, and even just perceiving life is a rather unfeasible task; one that even the most powerful artificial intelligence cannot begin to comprehend. That sparkle of life each and every being carries in them is unfathomable and perhaps no scientific methods would prove to be capable enough to attempt to put a frame around that sparkle.
Surely, everything changes and impossibilities do happen, just like once it was impossible to explore outer space, and before that, it was impossible to conceive that an entity like outer space existed at all… And we should all welcome change when it comes. Still, my being tells me that its eternity is a monumentality that is incomprehensible enough to be reproduced by zeros and ones.