An Archive at the End of the World
(Cheethams Library, Manchester, where Marx and Engels would study together. Photo taken by author)
At the end of the world, there will be an archive. All signs point to this. Long after we are gone, our data will be what remains of us. Perhaps my own legacy will be realized through whatever insights advertisers can wring out of my Tesco loyalty card and phone-derived GPS data. Beyond these more mundane innovations in data storage and archiving is the newer concern of generative Artificial Intelligence. This archive, then, is not merely the metadata corporations like to collect from us the moment we so much as breathe, but also constitutes millions of hours of creative work scraped off the internet to train AI.
Generative Artificial Intelligence itself functions like a massive repository, scraping the whole of the net as training material before it spits our own reflection back at us, perhaps in a half-distorted form. This is history at its most surreal, a vision of the world where forks have seven tines and women smile back with teeth that are just sharp enough to rouse your suspicion. If history is written by the victors, the vast archival work required to sustain Generative AI is itself an assertion of power by the capitalist class.
Understanding Generative AI as an archive might feel counterintuitive, given how future-facing the whole industry is. We are told by the likes of Elon Musk that AI is the bleeding edge of tomorrow, that it is the key to unlocking a new era of prosperity. The framing of AI as the “fourth industrial revolution” reinforces these changes not only as innovations but also as inevitable shifts in how we live. To object to AI is to possess the same sort of uneducated confidence as the original Luddites; it's to be left behind in the face of the inevitable.
Our relationship with technology, though, is as much about our past as it is about any ostensible future. This is particularly true in the case of Generative AI, which requires data on which to train itself. Because AI is only as good as the dataset you train it on, what data the software is trained on is crucial. In this sense, AI excavates the past to manufacture a new future in its image. The creations produced by the likes of Chat-GPT can be understood as combined outputs from the vast well of data such a system draws on every time you command it to make your grocery list or write you an essay on the merits of utilitarianism.
Here it might be helpful to take cues from Jacques Derrida, whose work fundamentally grapples with writing and language in relation to the reality they might construct. To Derrida, a central characteristic of an archive is violence. The archivist alone is the arbiter of what is remembered and what is forgotten, restricting what memory is by filtering out that which is deemed “irrelevant.” In this sense, Artificial Intelligence takes this filtration process and renders it on the largest possible scale. What is machine learning if not the process of filtering out incorrect or undesirable answers until something approaching reality is produced?
The ethics of AI and data scraping are murky. Programs like Chat-GPT use data scraped from the Common Crawl, an open repository of web crawl data. The consent behind this scraping exists in a grey area, with millions of artists having their work fed to LLMs without any consent given. Most notably, the New York Times is in the process of suing OpenAI for scraping the newspaper for the training of Chat GPT. The ethical consideration of an archive like Chat GPT, then, differs from the sorts of archives Derrida discusses.
In this sense, the archival violence inflicted by Artificial Intelligence differs from that of a typical archive because the information stored within an AI system is, for all intents and purposes, a black box. It’s an archive built for a particular purpose, but inherently never meant to be seen—it is the apotheosis of information-as-exchange-value, the final untethering of reality from sense. The opaqueness of this archive returns us to the initial question of capitalism without humans, of an archive without a reader, of form without content. When we are gone, is it this form of control that will remain our record of existence?
The archive at the end of the world articulates a capitalism without humans. Capitalism within this thought experiment is no longer a system of control or a manifestation of power but merely another set of cold gears, a piece of machinery that exists primarily to perpetuate itself. This mechanical understanding of the iron cage of bureaucracy isn’t unique—Weber and Kafka force us to ask if capitalism itself is capitalism without humans to exert power over each other or to provide their own labor to the cause.
In a sense, life within austerity already feels like a capitalism without humanity. Policy is stripped of all humanizing context, reduced to a few KPIs and metrics that must be reached at the expense of all else. Redundancies must be cut, indulgences put to the side. This system of bureaucracy is so vast and complicated that it basically cannot be fully understood by any individual. It is as if we live to sustain these institutions rather than the other way around. AI only further complicates what a genuine aesthetic challenge to capitalism might look like in a world that has already taken the option of surreality off the table.
We live in a world that no longer feels real. Baudrillard most famously discusses this notion of the “hyperreal,” but I struggled to find a cogent means of explaining the emotional nuance of living in a world populated by people who aren’t really people. Customer service agents are now chatbots that can’t be held accountable; I can’t help but wonder how long it is until the chatbots turn into an AI voice. Social media is populated with bots pretending to be people and people pretending to be bots. Craft fairs are being flooded with files printed off from midjourney.
Generative AI, the great archive at the end of history, can only spit out what we already have and know. This act of plagiarism and generation is more evocative of the cultural stagnation Jameson and Fisher describe than anything else in our current cultural landscape. There is nothing more hollow than a disruption borne of a million concurrent acts of plagiarism, the result of a final grasping at novelty in a world where even basic necessities are seen as unnecessary indulgences. The record of the past that Generative AI presents us with is inherently a surreal one. This is the past, but remixed and untethered from the reality it was scraped from. It is no wonder, then, that artistic techniques associated with the legacy avant-garde (surrealism, deconstruction) have lost their charm in a world that is more surreal than surreality itself
It becomes a challenge, then, to outsmart this archive. How can you be more surreal than surreality itself? How do we process all of this history when we seem to exist within a cultural landscape populated with pastiche and redos of greater movements in the years gone by? To play at this sort of future rupture is to play a losing game. Fisher is too caught up in a modernist notion of artistic evolution for his ideas to fully capture the reality of post-modern life. The object of aesthetic politics should diverge from that of surrealism because we will never be able to be more surreal than the archive itself.
The question is, of course, how might we imagine new futures or alternatives if we have given up on the avant-garde project altogether? The answer here is one I am unequipped to give as a mere master’s graduate. But I can think of the start of a solution, one which emphasises our own humanity in contrast to the cold mechanisation of the archive at the end of the world. If bureaucratic capitalism strips us of our humanity, then perhaps this is a viable starting point for a version of art which provides a genuine alternative to the reality which we face.
If we cannot outrun the end of history with the perpetual ruptures of a modernist avant garde, then perhaps the next logical course of action is to burrow under it, into the fertile soil of everyday life, where perhaps we will find the answers to the problems that brought us here in the first place.
