The term Artificial Intelligence (AI) has been around for many years, for most of which time it was more of an aspiration (or a foreboding, depending on your point of view) than an actuality. This began to change over the last couple of decades or so, when various kinds of AI systems started to establish themselves into our daily lives almost unnoticed. Much more recently, in considerably less than a decade, so-called generative AI has become so pervasive that people just call it AI. If you want to grasp the basics of how it works, take a look at Chapter 1 of René Daumal’s novel Mount Analogue (left unfinished at his death in 1944). When the narrator and his new friend Pére Sogol take a walk through Sogol’s indoor ‘park’, he tells us that, ‘Along the path, glued to the window panes or hung on the bushes or dangling from the ceiling, so that all free space was put to maximum use, hundreds of little placards were displayed. Each one carried a drawing, a photograph or an inscription, and the whole constituted a veritable encyclopaedia of what we call human knowledge. … I began to understand the use to which he put all those pieces of cardboard which spread out before us the knowledge of the century. All of us keep a fairly extensive collection of such diagrams and inscriptions in our heads; and we imagine we are “thinking” the loftiest scientific and philosophical thoughts when, by chance, a few of them fall into a pattern that seems neither too conventional nor too novel. It happens by the effect of drafts or crosscurrents or simply by the result of their own constant shiftings, like the Brownian movement, which displaces particles suspended in a liquid.‘ If you replace Sogol’s placards with ‘everything on the Internet’ and the ‘drafts or crosscurrents’ with ‘sophisticated algorithms running in gigantic data centres (with, incidentally, a huge environmental impact),’ then you have a pretty good picture of how this technology works. Is it also how our thinking works?
One of the great themes of G I Gurdjieff’s book Beelzebub’s Tales to His Grandson is that what we call our thinking is indeed an unconscious mechanism. Depending on the material that is fed into it, it can become very persuasive, very good at saying the right thing and saying it well. It can even become adept at persuading and manipulating people by using flattery and generally playing on people’s weaknesses. It does all this unconsciously and mechanically, without compassion and without remorse. Which is not to say that it doesn’t know how to seem compassionate or remorseful, but the real impulses themselves are foreign to its nature, so that it cannot be expected to understand them. In a sense it can’t really be blamed, because it’s just a machine.
Real thinking, he tells us, is something of an entirely different order. It requires a harmonious participation of the animal intelligence in our bodies and, crucially, of the life of the emotions, so that whatever one thinks is also simultaneously felt and sensed. But this is a vanishingly rare occurrence, because these different aspects of our make-up have never learned to understand each other. It’s so rare, in fact, that we don’t even recognise its possibility, and believe that this mechanical pseudo-thinking is all there is.
This tragic situation is memorably summed up in the very last chapter of the book, when our psychic organisation is compared to a Hackney carriage, with our so-called thinking as the feckless and irresponsible driver, unable and unwilling to care for the abused and neglected horse (the feelings) or the carriage itself (the body). What makes the situation tragic is that, even if somewhere in ourselves we long for this possibility of real thinking, our pseudo-thinking doesn’t recognise this longing for what it is and doesn’t know what to do with it.
Is there a way out? Beelzebub’s Tales gives clear indications that there is, but I’m reluctant to add more grist to the AI mill by saying more about it here. All the words we need are already widely available. But they will remain just words if we read them in the same old way, interesting ideas to agree with or disagree with. We need to feel that longing.