Artificial Intelligence Doesn't Have a Right Hemisphere
Language is the prime material of logical constructions. Owen Barfield, Poetic Diction (1928) Large language models are trained on human data from the Internet. Lex Fridman, #370, Edward Frenkel
Large language models: The fuel of today's artificial intelligence.
Artificial intelligence gathers the Internet's vast sea of words, processes them, and arranges them in a manner that allows it to accomplish tasks. That's how a large language model works.
LLMs are machine learning models that utilize deep learning algorithms to process and understand language. They’re trained with immense amounts of data to learn language patterns so they can perform tasks. Those tasks can range from translating texts to responding in chatbot conversations—basically anything that requires language analysis of some sort. Fast Company
Language, language, and more language.
But reality is more than language. The Reality Spectrum is:
The Tao ----------------------—>Essence---------------------—>Existence
Only the areas of essence and existence are subject to language. The ultimate reality--the prior reality--of the Tao isn't.
The first line of the Tao Teh Ching tells us that
“The Tao that can be put into words is not the real Tao.”
The Zen master, Wen-yu, echoed Lao Tzu when asked to explain the First Principle of Zen: “If words could tell you, it would become the Second Principle.”
At the end of his life, Thomas Aquinas had a vision of ultimate reality and quit writing, telling his friend Reginald that all the words he had written were mere straw.
If artificial intelligence is an accumulation, manipulation, sorting, reorganization, and re-presentation of words, it's limited. It is incomplete. It lacks the realm of reality that transcends words: the Tao.
That's a major desideratum. Without the Tao (no intuition of it, appreciation for it, or method of connecting to it), artificial intelligence can never be human intelligence. It can never replace humans. If it's limited to what is known and familiar, it can't invent and imagine.
AI Might Destroy Us but It Can't Become Us
But that doesn't mean it can't cause major problems. Indeed, it could wipe out humanity. It could, for instance, be programmed to write code that would implement a program to prevent garden tools from rusting. It would realize oxygen is a major cause of rust and implement a program to eliminate oxygen from the atmosphere (that's roughly the example provided by Max Tegmark at MIT).
AI is scary.
But AI cannot be human.
AI is the Left Hemisphere's Latest Roguery
Do you know what else is scary?
A left hemisphere unhinged from the right hemisphere.
A rogue left hemisphere that overwhelms its right hemisphere master is a terrible ruler. In Nietzsche's fable, the kingdom crumbles when the emissary usurps the master's role.
Modernity is the left hemisphere gone rogue. Artificial intelligence is just its most potent manifestation, as evidenced by AI's fundamental reliance on language.
AI's reliance on language is instructive and troubling.
Language is a tool of logic and reason. It is the “prime material of logical constructions,” to borrow a phrase from Owen Barfield’s 1926 classic, Poetic Diction.
Iain McGilchrist in The Master and his Emissary calls language “a means of manipulating the world” (113). He doesn’t mince words:
[L]anguage’s role is in giving command over the world. . . Words alone make concepts more stable and available to memory. Naming things gives us power over them, so that we can use them . . . [C]ategory formation provides clearer boundaries to the landscape of the world, giving a certain view of it greater solidity and permanence. . . Language hugely expands the range of reference of thought, and expands the capacity for planning and manipulation.
Language is what we use to capture. It is a source of power: a way of making our way through the world.
Language, in other words, is a great good, but it's not sufficient and, indeed, isn't remotely the most important thing. It's even questionable whether language is necessary for knowledge.
But to the left hemisphere, language is crucial. Language is its fuel. The left hemisphere places primordial interest in language, even though language is, as Jacques Derrida showed, manifestly limited and inadequate to capture full reality. Indeed, most thinking doesn't even involve language. We should not be deceived, McGilchrist observes:
into believing that language is necessary for thought. It could even be an impediment to it. Most forms of imagination, for example, or of innovation, intuitive problem solving, spiritual thinking or artistic creativity require us to transcend language . . . Most thinking, like most communication, goes on without language. Master, 107
Ultimate reality is paradoxical and paradox shows the "necessary limitation of our customary modes of language and thought." Master, 200. When confronted with paradox, with something that logic and language can't parse, it's stymied. The left hemisphere is a paradox moron.
It's a major problem. Lex Fridman is beginning to perceive this, commenting on a recent episode that it's beginning to appear that "Paradoxes are actually fundamental to reality. . . . We exist in a world, of not forms, but of paradoxes." Lex Fridman (Edward Frenkel, #370), 1:10:25.
If ultimate reality transcends language, what does that say about a left hemisphere that relies on language to function?
What does it say about a modern world that is underwritten by the left hemisphere?
What does it say about an artificial intelligence that, despite its awesomeness, relies exclusively on mountains of words?
It says that the left hemisphere is terribly limited and that our modern world is terribly limited because it is molded by the left hemisphere and that artificial intelligence is existentially limited . . . dangerous, but hopelessly and existentially limited.
The Right Hemisphere Transcends Language . . . and, Therefore, Transcends AI
The right hemisphere stands outside of language. It works with the left hemisphere, and in healthy people, mediates and instructs the left hemisphere, but unlike the left hemisphere, it isn't limited to language.
Because it does not rely on language, the right hemisphere is able to connect to the Tao (or maybe its ability to connect to the Tao allows it to transcend language . . . I'm not sure).
I earlier called the Tao a transcendental router. I'm rethinking that metaphor. I think it's more like a transcendental modem. Transcendence (the Internet) connects to the Tao (the modem), which in turn connects to the right hemisphere (the router). The right hemisphere then sends the Tao's signal to the left hemisphere (the digital devices) for implementation in worldly affairs.
It's the right hemisphere and its ability to accept (heck, laugh with) paradox that is crucial to comprehend fully reality, including transcendence. It's why current attempts to use AI as spiritual counselors are ridiculous.
If a Person Has a Functioning Right Hemisphere, He Will Recoil at Attempts to Use AI in Right Hemisphere Areas
I think people intuitively realize this. I hope every sane person intuitively recoils at the idea that robots can give spiritual advice.
But when I say "sane," I'm referring to people who still have a right hemisphere that meaningfully influences their mental world.
I suppose that, if a person has hopelessly squashed his right hemisphere through modern living, he might not intuitively recoil. It's hard to say. It's even hard to imagine, but there are apparently people who embrace the possibility, and it's disturbing.
It might be "Exhibit A" on the list of a modern world that has lost all contact with the full Reality Spectrum.
Which would make it "Exhibit A" in a brief about how the modern world has gone mad.
The Right Hemisphere is Our Hope Against AI
Only the ri
ght hemisphere can get us out of the artificial intelligence problem: that technological, governmental, and capitalistic efforts to make the most money and become the most powerful will destroy all of us.
Max Tegmark calls it "Moloch."
He has also called for a six-month moratorium on AI development until we can figure out what exactly we're dealing with, perhaps giving governments a chance to control it.
I think a six-month moratorium is a good idea.
But I think we're just stalling.
Central governments today are dominated by Moloch . . . which is a demon driven by the left hemisphere, just as Carthage was (more on that later).
To rely on a left hemisphere machine to stop another left hemisphere machine?
It's just more madness and merely kicks the existential can down the road.
So what would I recommend? It's a fair question, but I don't have an answer. We need another Solon . . . or an army of Solons. We need politicians and an aristocracy with robust right hemispheres, but if you think today's world leaders have robust right hemispheres, I'm afraid you're the kind of person who asks a robot for spiritual advice.