AI-anxiety is all over the news and on our minds. What can we do about it?
Is AI taking over? YES, argues Douglas Hofstadter, a renowned scholar of cognitive science, physics, AI, and comparative literature. A longtime skeptic about AI’s capacity to become humanlike, Hofstadter recently changed his mind. David Brooks just published a NYT opinion piece, citing Hofstadter: “Human Beings Are Soon Going to Be Eclipsed.” Are we playing with fire - and about to lose the human race?
AI anxiety is warranted but misplaced. If the crux of the matter is thinking, which equals consciousness, as Hofstadter argues, we lost an edge we never had. Many animals, including mammals, birds, and at least some cephalopod molluscs (octopuses, squid, cuttlefish) seem to feature consciousness (Ben-Haim et al. 2021; Birch et al. 2020).
Brooks writes he wants to build a wall around the “sacred… essence of being human.” This is understandable. We want to keep what’s ours, otherwise, who are we? But ethology has already brought down that sanctifying wall. Many animals are remarkably intelligent, sentient, and emotional. Decoding humanness may require that we go beyond sacred essence.
The crux of the human matter, and mind, might be not thinking and consciousness but self-consciousness. Humans are capable not only of witnessing the world but themselves, engaging in self-reflection that often involves inner dialogue (Fernyhough 2016). Hannah Arendt (1971) went further to demand that thinking involve a critical inner dialogue, but knowing people, we may lower the moral bar. There is no essence to being human, but cultivating open inner dialogue is essential to realizing our humanity.
What then about our relatives? Though it’s obviously impossible to know for sure – how can we ever know “what it’s like to be a bat?” (Carruthers 2019; Nagel 1974) – non-human animals likely lack self-reflection. Few animals do pass the mirror test, but probably don’t converse with themselves. One morbid indication is that while animals suffer and fear pain, and some may even grasp the finality of death, they never commit suicide (Humphrey 2018).
What about AI? We may as well return to Camus’s (1955 [1942]) arresting opening of The Myth of Sisyphus: “There is but one truly serious philosophical problem and that is suicide. Judging whether life is or is not worth living amounts to answering the fundamental question of philosophy.” The time may have come to move from the Turing test to the Hamlet test: Can A(G)I originate a soliloquy? Can it be suicidal?
A lead might be Julian Jaynes’s (1990 [1976]) unruly thesis about The Origin of Consciousness in the Breakdown of the Bicameral Mind: We used to think that the voice in our head was God’s, but when we realized its our own, something new was born.
Something new was born, and, we might add, God died. As Nietzsche guessed, the signs of humans’ failing to supersede Him are all over the place – to fight (harder) to replace God, or flight from our humanity: becoming machinelike, or substituting doglike for godlike aspirations, basking in “man to man a wolf.”
A seeming side effect that goes to the heart of it is discarding our inner-dialogue. A couple of years ago when I asked my students about their own inner dialogues, my classes were roughly equally divided: some students took this internal conversation for granted, others had no idea what the rest of us were talking about. Increasingly, the balance is tilted toward the latter group. Maybe we’re rebuilding our species’ Bicameral Mind, only this time assigning God’s role to AI, finally realizing deus ex machina.
The title of Brook’s piece is intriguing: “Human Beings Are Soon Going to Be Eclipsed.” The celestial dynamics of eclipse go against my intuition, read hubris. I assumed that when Earth gets in between the sun and the moon, it casts its shadow on the latter. I was wrong. The earth is not that big, the moon not that small, and both slightly tilted. So this three-body alignment typically leads to the exact opposite: we get our monthly full moon, rather than eclipses, lunar or solar, which are infrequent, and rarely total. Maybe Hofstadter is right, and we’re getting one of those total eclipses right now. But if we’re going to go Godly, we might as well learn from His sticky note in the sky, drawing the rainbow to remind Himself to never destroy all living (Genesis 9, 12-17). Perhaps we should look at our sky to be a bit humbler, to see the value of a tiny tilt, and to recall: Stars never eclipse; what does eclipse, AI included, never shines.
References
Arendt, Hannah. 1971. "Thinking and Moral Considerations: A Lecture." Social Research 38 (3):417-446.
Ben-Haim, Moshe Shay, Olga Dal Monte, Nicholas A. Fagan, Yarrow Dunham, Ran R. Hassin, Steve W. C. Chang, and Laurie R. Santos. 2021. "Disentangling Perceptual Awareness from Nonconscious Processing in Rhesus Monkeys " Proceedings of the National Academy of Sciences 118 (15):e2017543118.
Birch, Jonathan, Alexandra K. Schnell, and Nicola S. Clayton. 2020. "Dimensions of Animal Consciousness." Trends in Cognitive Sciences 24 (10):789-801.
Camus, Albert. 1955 [1942]. The Myth of Sisyphus, and Other Essays. 1st American ed. New York,: Knopf.
Carruthers, Peter. 2019. Human and Animal Minds : The Consciousness Questions Laid to Rest. First edition ed. Oxford ; New York, NY: Oxford University Press.
Fernyhough, Charles. 2016. The Voices Within: The History and Science of How We Talk to Ourselves. New York: Basic Books.
Humphrey, Nicholas. 2018. "The Lure of Death: Suicide and Human Evolution." Philosophical transactions of the Royal Society of London. Series B, Biological sciences 373 (1754):20170269.
Jaynes, Julian. 1990 [1976]. The Origin of Consciousness in the Breakdown of the Bicameral Mind. Boston: Houghton Mifflin.
Nagel, Thomas. 1974. "What Is It Like to Be a Bat?" The Philosophical Review 83 (4):435-450.
Postscript
I had a brief chat with Chat-GPT to see if it's about to pass the Hamlet Test. It ain't.