The obvious thing for this article would have been to write it using ChatGPT, the new AI device capable of generating synthesised writing indistinguishable from that of a human author. ChatGPT isn’t like earlier, clunkier composition programs, which proved incapable of generating plausible copy for even the most basic filler. Instead, thanks to the exponential expansion of computing power, ChatGPT can take in and ‘learn’ from all text on the internet, and not only spit out writing that is naturalistic, but write in particular styles, and mimic certain authors and genres. It finds facts, integrates them, makes arguments. It responds to requests phrased in the same informal way you would use talking to a human writer. A recent New York Times piece by a US congressman and former software developer began with a sizzling opening paragraph full of expertise and foreboding, that was written by ChatGPT to be a sizzling opening paragraph in the New York Times op-ed style. Anyone who suggests that they have an easy answer to what this represents is lying, not least to themselves.
The truth is that a whole series of professions and disciplines are freaked out by this, more than anything that has previously come on the scene. Academics, writers and editors aren’t just disturbed by this technology; they are petrified by it, if they are honest. Universities have scrambled to respond, some by resorting to pen and paper exams, others by allowing ChatGPT’s use—a decision which could be described as ‘courageous, Minister’. Newsrooms and journalists’ unions are scrambling to work out what to do, all the while dealing with the unnerving feeling that their whole mode of being and acting on the world has suddenly been superseded. Many thousands of jobs that had developed in the world of ‘copywriting’, which mostly involve providing filler for the internet, will now disappear.
But these social and economic changes are relatively mundane, and their status as repetition—how many crafts have been undermined in history, how many classes of job lost?—conceals the radical shift that ChatGPT represents in crossing the barrier of the Turing Test less than a century, one long lifetime, after it was first proposed. The Test famously argues that a computer can be said to display ‘intelligent’ behaviour when a human, in extended text-based interchange with it, cannot tell if they are engaging with another human or a machine. Various bits and bobs of tech have been able to win this ‘imitation game’ hitherto. ChatGPT smashes it absolutely.
The reaction to this development has been largely technocratic, instrumental and technophiliac, even, or especially, from what remains of the critical Left. Much of the discussion of ChatGPT’s incidental effects strikes me as primarily a way of avoiding the deep challenge it presents. The challenge itself is once again being constructed around the warning throat-growl of ‘Luddism’: to ask questions as to whether Silicon Valley capital should be able to determine the technological transformation of our life world is presented as being against the horse-bridle.
There is a circularity in this process that loops through decades of the history of modernity. That there is a fundamental institutional inability to take a critical, reflective grasp of a new technology such as ChatGPT—and AI visual composers such as Dall-E and Midjourney—is a product of the victory of a manner of thinking that has made its runaway development possible in the first place. This is the ‘analytic’ mode of thinking, born in Vienna and largely Anglo-American thereafter, which responded to the European reflective tradition with a denial of its holistic and interpretive orientations.
Turing’s development of modern computing took place within a framework the thought-practice of which was oriented to breaking down any distinction between surface and depth in reflection on human subjectivity and being in the world. He was part of a world in which Wittgenstein and the Vienna Circle had reduced truth to statements about sense-data and tautologies. Wittgenstein’s later philosophy then reduced language to its uses and philosophies to the disentanglement of everyday language practices. Gilbert Ryle derived a behaviourism from his notion of the ‘category error’ of levels of being, his example being that of someone showing a visitor round a university and pointing out buildings, a library and so on, only to be asked, ‘But where is the university?’ Ryle took this to be an indication not that things like universities, nations, and minds were real in a manner other than that of apples and stars, but that they were not real at all. In that conspectus, intelligence becomes the necessary supposition arising from multiple manifestations of its effects.
The invention and spread of the computer has been a product of this intellectual deconstruction of purposeful human being in the world, made possible by Claude Shannon’s insight that electric circuits matched the graphical representation of formal logic reasoning. Once a certain level of computing power had been reached—coinciding by that very fact with the creation of the World Wide Web as a top layer of the internet—the ‘intelligence-imitation’ drive took over, and now has simply become ‘the internet’. With such analytic processes made real by the application of trillions of dollars, the philosophical approach has become not observational but programmatic: a world flooded by non-physical machines carried by physical ones, which are replicant processes ascending all the way to full discursive engagement. This is a world in which the critical, reflective, purposeful subjectivity and the centrality of full reciprocal human exchange is undermined by the machines made to serve and extend it. With such developments proceeding unrestrained, we will become more like these imitations of ourselves than they will become like us.
The reason that this cannot be grasped in public debate—especially in the Anglosphere—because of that old habit of modern intellectuals which McLuhan identified as ‘Gutenbergian’ thinking: the idea that the world, intellectual and otherwise, created by mass paper printing was the ‘natural’ one, and that all before it was myth and all after it fad. This remains the default position of right-wing op-ed writers around NewsCorp, for whom Twitter and Facebook are the intelligence-less Fall from the world of plain common sense while the 800-word op-ed generated for a newspaper, and from the intellectual framework print brings into being, is a natural communion with truth.
The challenge that ChatGPT represents can only be understood if it is seen as being part of a process of ‘categorical’ transformations that began with writing itself—and if we take from that not the belief that nothing potentially catastrophic has happened in this process, but that something potentially catastrophic has always been happening. If all these technologies, as they appear, stand humanity back from its given conditions, its materially abstract social relations and the character of the object-world, then the successive reiteration of such transformations, if unreflected upon, promises to reduce reciprocal purposeful human relations to a mere trace within a system operating autonomously. The science-fictive possibilities—evil machines killing humanity with gas bombs and so on—are in some way less terrifying than the prospect that the ‘ahumanising’ orientation of technological development can take a species whose life is grounded in meaningful collective purpose, attachment and love to a point on the ‘other side’ of such and deprive it of the collective capacity to create a full understanding of this process, materialising instead a thin reconstruction of such relations for a significant stretch of historical time. That the arrow of technology and materialised social abstraction is also the means by which a human liberation from the natural conditions in which we arise as a natural-and-cultural species is possible has been so often said that it doesn’t need much more than namechecking. The technophiliacs of Right and Left sound like Borges once said Argentinian pro-Nazis did as the Blitzkrieg possessed: as terrified as he was, despite their triumphant celebration. Right-wing, ultra-neoliberal accelerationist types are at least consistent: they have no idea of ‘the good’ other than the quantitative advance of technology and transformation, its sheer velocity—an approach which makes things easiest by reversing qualitative change into quantitative. Such new technological challenges do not, on the other hand, make Marxism, which does have an idea of ‘the good’, into an inadequate framework to answer the challenges we face. They simply reveal that it always was one.
The appearance of ChatGPT might, had the tape of history—and particularly intellectual history—run another way, have been the occasion for a deepened discussion of what sort of relationship we want to have to technology. But any discussion of that sort, in any field, has to work as much backwards through existing technologies as it does forward through new ones. Consideration of ChatGPT in schools is not merely an occasion for asking about the dominance of computers in schools, but one for asking about the dominant role writing itself plays in education as opposed to making, doing, dialogue and presence—a form of schooling that would come out of the classroom completely. And one can say the same for life in general. It is the dominance of the ‘arrow’ of writing, printing, the digital world and the institutions and regimes they make possible, from the nation to capitalism, that the world needs to step back from, but which intellectuals, now near-wholly constituted by such practices in ever-larger areas of their life, find hardest to perceive.
With our capacity of critical reflection ‘stuck’ like this, foreboding about any technology will always seem Luddite, and will be constructed as the very thing which one must overcome. For forty years or so, in the wake of the Second World War, that capacity appeared in the world in a mass form as the ‘New Left’. Its success was in part due to the fact that the war introduced the critical, reflective and synthetic thought processes of Europe to the realm of disassembled practical and intellectual parts that reconstituted the United States. That it has now broken up shows that the sheer volume and transformative character of new technologies does not, of itself, generate the mass critical and reflective capacity needed to enable the fully human prospect to reappear. That gap—an insufficiency that may prove fatal—was the occasion for Heidegger’s malignant, despairing cry (even during the New Left) that ‘only a God can save us now’. That cry cannot be accepted by anyone whose entry to a ‘politics of being’ began on the Left and is oriented to and by the possibility that a mass critical capacity with regard to technology will itself become a historical force within human life. Which is all one can work towards. At the moment, however, prior to any conversions between those who understand the challenge but lack the ways to think about this, it is almost a full-time pursuit to continue to think clearly about these things oneself. Whoever or whatever wrote this.
Zero Gravity: Floating towards posthumanism
Richard King, Sep 2022
The transition to the posthuman will not turn on any Judgment Day scenario, but will be a slow and incremental process of quantity turning into quality.
Comments
Reading this insinuates sensibility into the suicide of David Foster Wallace and the intense magnetic field of Beckett’s spotlighting of Lucky warming to his task in ‘Waiting for Godot’ . Neither is a ‘consummation devoutly to be wished’, but Guy’s ‘materially abstract social relations’ are definitely fluctuations within those kinds of event horizons. In his conspectus, necessary suppositions appear superfluous to the manifestations of his intelligence, and are of necessity, unnecessary. The moving finger writes, and having writ, moves on until it strikes a halting problem, as it did for Guy’s terminal rhetoric ghosting Turing’s test.
I feel that the breathless learnedness, literary allusiveness, olympian detachment and eloquence of this grandiloquent and condescending ‘writing effect’ is somewhat undermined by the fact that it appears in the comment section.
I think there’s some significance between the hand-coded authoritarian control logic of mid-20th century symbolic AI (computing), and the statistically-based hyper-dimensional perceptual models of neural networks.
I have the freedom to ask a neural network for anything, as long as it’s within the confines of the market. And I’ll have absolutely no idea how it was made, but it will look very pretty.