Artificial intelligence, academic research and the movement of thought

 

The more I think of it, the more I believe that the role of academics and researchers cannot be to produce perfectly defensible claims, so carefully justified and qualified that there is nothing left to say about them. There is something strange about making this the ambition of thought. We write about living, changing realities, yet seem to want our own thinking to arrive at a form which no longer needs to move.

I've played around with AI long enough to recognize a certain pattern in the way it writes. Everything is measured. Every claim comes with its little protection against possible criticism. One reads and says: “Yep, that's about right.” And then what? Often, for me, nothing. There is no particular desire to respond, to oppose, to complement, to associate the idea with something else. It feels as though the thinking has already been done, and all that is left is to acknowledge that the answer is reasonable.

I believe that an idea can be wrong and still be worth thinking. More than that, its value may lie in what becomes possible through thinking it, including possibilities that would never have emerged if we had insisted on establishing its correctness first. The question for me is whether it allows individuation to continue. This is the criterion I want to explore, even if I cannot yet turn it into a neat set of rules for evaluating a paper in a peer-reviewed process.

Obviously, this doesn't mean that everything which crosses our mind deserves to be published, or that every exciting claim is valuable. Propaganda can be creative. Misinformation can stimulate enormous amounts of activity. But activity alone tells us very little. People can spend their lives repeating the same argument, becoming more and more invested in it, without ever allowing it or themselves to change. There is a lot of movement on the surface, while the structure of thought remains firmly in place.

This is where Simondon helps me articulate what I mean. He describes life as a continuing activity of individuation. A living being carries potentials that are not exhausted in its present form; it can resolve problems or tensions it encounters through its interaction with his milieu, via changes in its own structure or via acting on and altering the milieu, or perhaps, as is often the case, via both simultaneously. I want to take this further as a way of valuing thought. For me, individuation is its own reason. There need not be a final, perfect state which justifies everything that happened along the way. What matters is that the process can continue, that resolving a tension leaves room for further becoming.

A “perfect” biological form, impervious to every change or challenge, seems a strange ideal for a living being. And yet this is what some noetic or academic exercises feel like: attempts to build a thought which nothing can disturb. Every tension has already been answered, every objection has already been given its appropriate place. I don't mean that these texts actually achieve such perfection. AI can be wrong while sounding exceptionally reasonable. I'm describing the impression they give, and what that impression does to my desire to think.

So yes, I would value thinking incorrectly, provided that it doesn't undermine the possibility of further individuation. If I believe I can fly and jump out of a high window, the consequences may abruptly end my biological and noetic individuation. Reality has something to say about my theory, whether I find it stimulating or innovative or not. This is an obvious case. Other cases are much less immediate, and this is where I think the question becomes interesting.

Take the following thought experiment: imagine communities in conflict, each appealing to different divine powers, with religious authorities giving incompatible interpretations of what those powers demand. Now imagine the idea of a single god and a holy book becoming a shared point of reference, with a corresponding institutional hierarchical structure for interpreting it. It might make it possible for these communities to recognize obligations towards one another. Suppose, in this case, that the idea helps people move beyond a conflict they could not otherwise resolve. Would its value disappear if its account of the divine turned out to be false?

I don't think so. It may have allowed a collective life and experience to develop which was previously impossible. People can live together, establish relationships, think about things other than the conflict which consumed them. The idea has created enough stability for something else to happen. But eventually that same framework may become inadequate. Who speaks for this one god? What happens to someone who questions the interpretation? A way of overcoming an earlier conflict can itself become a source of tensions which it cannot resolve on its own terms.

In the imagined case, people might turn towards atheism, or towards scientific explanations and institutions which no longer require agreement about gods. They might also transform their religion from within. I don't need these possibilities to form an inevitable historical sequence. What interests me is how a system can make further life and thought possible, then become something that further life and thought have to work through. The fact that it is subsequently challenged does not mean that it should never have existed.

Nor would harmony be sufficient on its own. If everyone agrees because dissenters have been silenced, whose individuation is continuing? A community might appear stable precisely because some of its members have lost the possibility of contributing to its transformation. This is a real difficulty for my criterion. I cannot judge an idea solely by how much it allows its defenders to develop. I also have to consider what it does to the people who encounter it, live under it, or try to respond to it.

 

Coming back to my previous biological example, it would sound ridiculous to say that the biological “form” of an animal, say a Cephalogale, the ancestor of our current bears living some 30 million years ago, is “less evolved” than the bear. On the contrary, biology or life is always perfect, in the sense that its current biological form is the reflection of the current tensions and problematics existing between itself and its milieu. But when we consider the development of human thought, we apply formal “true/false” and “good/bad” judgement upon it rather than interpreting it as adapted to the current coupling of humans to their shared natural as well as to their noetic milieu. What counts, therefore, is not criteria of “truthfulness” or “veracity”, or “goodness”, but adequacy. A thought, just as a current biological form or structure, is adequate depending on its ability to metastabilize existing tensions and problematics between individuating beings and their milieu, be it a natural/physical milieu (in the case of biology) or a complex noetic, sociotechnical and technological milieu (in case of human beings). 

To give a noetic example, think also of Aristotle's explanation of falling bodies. He understood earthy bodies as moving towards their natural place at the centre of the cosmos. We no longer accept this as an explanation of gravity. Yet he developed an account that could be encountered, examined, and opposed. Should the work never have existed because it was wrong? I cannot prove that Newton would not have developed his theory without it, but I don't need to. My question is whether working through a mistaken explanation can help us formulate something we were not yet capable of thinking. I believe it can.

This is why I want academics to defend ideas that are substantiated enough to be taken seriously, yet still leave something open. A poorly argued novel theory often leads to quick closure: I can dismiss it without much effort. A well-argued theory which goes against my understanding forces me to do something. I may disagree, but now I have to work out why. And in doing so I may find that my own understanding was less complete than I thought. Even if I never accept the theory, I have moved.

This is what I mean by an individuation of thought through the encounter with a noetic milieu: the environment of ideas, arguments, books, and people with which we think. Something does not fit between the way I understand things and what I am encountering. I have to create a relation, revise an assumption, perhaps discover a question I had not seen. I want writing which makes such encounters possible. An engaged person defending an imperfect idea often does this for me. Their insistence makes me want to answer them. There is something there to take hold of.

AI is useful to me when it helps this movement continue. It can deepen an intuition or show me where an argument doesn't hold together. But I don't want every unusual claim removed simply because it is the easiest part to criticize. Sometimes that is exactly where the thought is happening. Help me work on it. Show me what it implies. Ask me why I insist on it. There is a difference between discovering that I need to abandon an idea and having it quietly edited out before I have understood what I was trying to think.

And I don't want to make “human” and “AI” into fixed categories which settle the matter in advance either. If an AI response moves my thinking, then it moves my thinking. The answers from an AI should become something to think with, and sometimes against, rather than the point at which thinking is supposed to stop.

Some AI writing still gives me the image of a perfectly calm, infinite ocean, reflecting the sky. There is no texture, no movement, just an almost unnerving stillness. I can appreciate how well everything fits together, but I don't feel drawn into it. I would rather read something that leaves me with a thought I cannot quite put down, like a breeze and waves that carry a boat to an unknown destination. Perhaps I will discover that it was wrong. Perhaps, in trying to say why, I will discover something else. I don't want that possibility polished away.

Perhaps in the future, humans will come to view noesis and thought itself as an artistic endeavour, across entire fields including academia, rather than setting strict boundaries for novel thinking, reserved for science-fiction books and fantasy novels. Think of painting: should artistic painting be reserved to canvases whereas painting a room in a house be strictly utilitarian and methodic? These binaries can be broken down, and many artists have innovated in the way a room is painted to combine both art and utility. AI can become the ultimate “painter”. It can flawlessly paint all our rooms white. But when we find ourselves in these rooms, surrounded by perfect noetic symmetry, mathematically and geometrically calibrated, what effect will it have on the movements of our minds? Given AI’s tremendous advances, in light of the recent GPT6-Astra release, opening up novel criteria for judging the value of human thinking and writing seems more than justified. In the future, we should all become artists of thought, enhanced by AI, supported by it, but certainly not replaced by it. That is the kind of balance we will need to find in the future.

by Martin Schmalzried , AAIH Insights – Editorial Writer