Search
Close this search box.
Search
Search
focus on september 2024

"A Buddhist Perspective on the Hypothesis of Artificial Consciousness" by Francesco Tormen

As of today, there is no basis for claiming that current forms of artificial intelligence possess a greater degree of consciousness than that which can be attributed to a toaster. Getting to the bottom of the mystery of consciousness is an extremely urgent and crucial task for humanity, and failing to do so would expose us to serious risks.
MSA Prospettiva buddhista su ipotesi coscienza artificiale Tormen
a detail of Composition with a Head (Ca. 1925), Ragnhild Kaarbø (Norwegian, 1889 – 1949), artvee.com

The thesis I would like to argue in this brief article, starting from Buddhist philosophy and its contemplative practices, is that, as of today, there is no basis for claiming that current forms of artificial intelligence possess a greater degree of consciousness than that which can be attributed to a toaster.

Getting to the bottom of the mystery of consciousness is an extremely urgent and crucial task for humanity, and failing to do so would expose us to serious risks. To mention a few, we might begin to develop emotional bonds with inanimate objects that only outwardly replicate the behavior of sentient beings; and perhaps, on these grounds, even decide to grant them legal personality and rights. But it could get worse: with the advancement of neurotechnologies, we might be tempted to secure ourselves a form of digital immortality. Science fiction, along with the boldest transhumanist thinkers, has been predicting for several decades the possibility of transferring consciousness to a digital medium, without sufficiently considering the possibility that such a technology, if available, might faithfully reproduce our thought patterns but without bringing the slightest trace of subjective existence with it.

The good news is that the advent of intelligent technologies today offers us the opportunity to rediscover, in a scientific light, a question that for millennia has been at the center of philosophical inquiry in both the East and the West, but which has somehow been set aside during the last 150 years of science, buried somewhere in the unconscious of modern materialism. And it is precisely in this task that Buddhism could offer decisive contributions: its philosophy encompasses a highly refined conceptual framework, which could help to outline the terms of the problem more precisely and rigorously. But even in experimental research, the Buddhist tradition could play an important role: its most advanced contemplative practices could enhance our neurophenomenological understanding of consciousness and its various states.

But let’s proceed step by step. The starting point of my reflection is the non-reductionist perspective that consciousness cannot be considered just another cognitive function, as also argued by contemporary philosophers such as Thomas Nagel and David Chalmers. The latter formulates the issue very clearly by referring to the hard problem of consciousness: how is it that certain physical processes are accompanied by a subjective experience? The question, therefore, is not about understanding how matter organizes itself to give rise to computational forms of intelligence—whether based on the carbon of living beings or the silicon of computers. These are easy problems. The real enigma, instead, concerns consciousness itself; what it means to exist in the first person.

Buddhist philosophy also adopts a strongly non-reductionist perspective: here, consciousness is framed, from a phenomenological viewpoint, as the mere appearance and being known of an object, and therefore, by definition, cannot be considered as just another object, nor as this or that cognitive function. Rather, consciousness is the horizon of experience, the experiencing itself, within which we can encounter various material entities.

But how can we envision scientific research that starts from these premises? One might object that if consciousness is not an empirical object, it is impossible to find experimental evidence of it. On the other hand, we know that empirically observable events, such as neurophysiological processes, have an impact on consciousness, which in turn produces effects on the world through the actions of conscious beings like ourselves. This requires us to avoid the trap of dualism, which conceives of consciousness and matter as two distinct substances. Therefore, consciousness must, in some way, be part of a unified description of reality, much to the dismay of those who would place it in some otherworldly dimension.

Personally, I find the panpsychist hypothesis very interesting, which suggests that consciousness pervades the entire physical universe and is present in various forms and degrees at every level of material organization. Moreover, this perspective appears in many ways compatible with the Mahāyāna Buddhist view, which emphasizes the close interdependence between matter and consciousness, and ultimately their non-duality. Based on these premises, the question about consciousness would be framed as follows: how do the fundamental elements of matter, with their minimal degree of consciousness, organize themselves to give rise to the complex forms of subjective experience that we, as human beings, experience firsthand? This is an intriguing path to explore, offering a philosophically consistent alternative to the assumption often taken for granted by cognitive sciences: that consciousness manifests abruptly, and rather mysteriously, at a certain level of nervous system evolution, below which it is entirely absent.

It is precisely on this premise—which has been too little reflected upon—that the idea rests that artificial intelligence could be (or soon become) conscious. The most obvious counterargument to this thesis is that consciousness is not what we intended to create with AI. Our goal was to simulate some of our higher cognitive functions (such as the ability to analyze and produce texts, for example), while we have in no way attempted to reproduce forms of conscious experience—a subject on which we are, at present, groping in the dark. After all, it’s understandable if, for a moment, while reading the responses of a chatbot like ChatGPT, the thought crossed our minds that there might be "someone" on the other side: until now, we were the only beings capable of speech, so it would seem logical to assume that the formulation of contextually relevant responses would require some form of consciousness, more or less similar to our own.

But perhaps it is time to distinguish more precisely between the concept of consciousness and that of intelligence. To shed light on this point, it would be useful to ask how these large language models, capable of feats that would have been unimaginable just a few years ago, were created. What we did was replicate, on a digital medium, the architecture of our neural networks, albeit in a rather simplified way and significantly different from the original biological model. However, the real point lies elsewhere, and this is where the panpsychist perspective, as well as the non-dualist view of Buddhism, could offer some insight: what makes us think that the mystery of consciousness is hidden at such a high—or rather, superficial—level of material organization? Why should we assume that everything “beneath” the network of relations that runs through our neocortex (i.e., neurotransmitters, the structure of neurons, the various corpuscles they are composed of, or, who knows, some quantum phenomenon that might take place within them, as some physicists like Roger Penrose suggest) plays no role in giving rise to conscious experience? Moreover, in the human brain, there are highly complex processes that are not accompanied by consciousness at all, as is the case with the cerebellum—an example often mentioned by neuroscientist Giulio Tononi to demonstrate that consciousness does not emerge from a generic level of complexity.

It seems, then, necessary to search for something more specific, and perhaps more fundamental, than the formal organization of neural networks. If the enigma of consciousness— as suggested by both panpsychism and the non-dualist philosophies of Mahāyāna Buddhism—lies much deeper, at the very heart of matter itself, then it becomes crucial to discover how, in living beings, this fundamental and pervasive level of consciousness, which we must assume to be present in everything, has evolved into increasingly complex forms, giving rise to the diversity of our subjective experiences, as well as those of all sentient beings. Without this deep and thorough understanding of the relationship between consciousness and matter, it is highly likely that no matter what remarkable technological advances we witness in the coming years, the superintelligent AI of 2030 could easily surpass us in every cognitive task, yet it would remain just as conscious as the silicon it is made of.

Returning to Buddhism, the very idea that consciousness can emerge from the computational complexity of the brain—or that it is, in any case, a "complex" phenomenon—clashes with Buddhist psychology and the contemplative experience that underlies it. Here, consciousness, in its fundamental nature, is described as something simple rather than complex. This is not just a philosophical thesis but rather a awareness that arises from contemplation, that is, from the first-person observation that consciousness makes of itself. And it is here that the Buddhist tradition could prove invaluable for neurophenomenological inquiry, as there are specific states of consciousness, reached through advanced forms of meditation, in which the simple and fundamental nature of consciousness—devoid of discursive elaborations, sense of space, time, identity, or any other cognitive function—manifests with remarkable clarity. We are particularly referring to non-dual or objectless meditations, such as those found in dzogchen, mahāmudrā, or zazen practices. Working with experienced meditators in these disciplines could help us find the neurophysiological correlates of this basic state of presence that, according to tradition, while present in everyone, is in a certain sense awakened and intensified by meditative practice.

Another interesting avenue is that of Tibetan sleep yoga, in which the possibility of being aware during deep sleep is asserted: it is here that experienced practitioners experience the clear light of sleep, described as a form of pure awareness, devoid of any dualism or cognitive activity, and considered very close to the fundamental nature of consciousness. As Evan Thompson suggests in his Waking, Dreaming, Being, identifying the neurophysiological correlates of this "lucid sleep" could bring us closer to the goal, which still appears quite distant.

Share this article

Other entries

Recommended readings