Make it new? Art and knowledge in the age of automated content generation

The recent advent of ChatGPT, Midjourney and other content generators has provoked much hand-wringing over the future of creative production in the age of automated art. What these technologies might mean for both art and audience is very much up in the air, though many commentators have imagined that AI is likely to produce a profound change to our very means of making knowledge and artistic works.

I want to begin by suggesting that AI art intensifies already-existing tendencies within our contemporary economic model of platform capitalism rather than creating new ones. Under the auspices of late capitalist production processes, there has already been a substantial standardisation of art manifested in production, distribution and consumption—one that AI promises to exacerbate.

In the sphere of production, the professionalisation of the arts has over the past few decades of neoliberalism removed much of the unpredictability of cultural work. To work in the arts requires a level of financial independence unimaginable in the post-war economy, where the supports of the social welfare state helped produce a flourishing of creativity. The widespread funding of arts and humanities education for working people created a diversity of perspectives—from the art school tradition that produced such luminaries as John Lennon, David Bowie and Brian Eno in the UK to working class writers like Christos Tsiolkas and Tim Winton in Australia.

This relationship between state and art most amusingly gave rise to the rhyming slang for the dole, the rock n roll. But as the neoliberal ethos of user-pays education and austerity measures have reduced the social welfare state to a pale shadow of itself, and in the process narrowed the field of cultural production to only those few with enough social support to afford it.

Standardisation is manifested, too, in the methods of distribution—the algorithms that mediate our access to content in negotiating the vast databases (Spotify, Kindle, Netflix, etc) we call streaming. The infinitesimal distinctions of micro-genres from ‘Scandi crime drama’ to ‘lo-fi beats to chill to’ purport to provide us with the most niche of pleasures. However—as anyone who has spent any time frustratedly searching for something to watch knows—the proliferation of these categories ultimately relativises everything into only one form: content. Art has become ambient, far from the singular artwork that characterised it in its modern form—namely, the masterpiece.

Fredric Jameson several decades ago suggested that postmodernism involved the ‘unparalleled standardization of everything–feelings along with consumer goods, language along with built space.’ At its very beginnings, Jameson anticipated the way that the internet would swallow the field of culture whole, producing a globalised field of transnational production and consumption in which we by and large have access to the same texts which we then discuss in the same ways on the same social media in the flattened language of a global bastardised version of Cultural Studies inflected by US diversity discourse, or through the same semiotics of emojis, gifs, reacts—whatever.


What this standardisation has meant is that art has become stripped of its alterity and specificity. What Walter Benjamin called the ‘aura’ of the artwork has dissipated. Shorn of a specificity of time and space, it withers in the face of the proliferation of mechanically produced copies.

The movement from singularity and specificity, from the artwork at which I gaze to the opera I hear performed in front of me, has had a long and arduous journey across the twentieth-century era of mass communication to the twenty-first-century platform economy.

Enter AI generators, which remove not only specificity and singularity, but the very fact of the artist themselves. The death of the author that Roland Barthes critically heralded becomes intensified by the new power of the user-artist to generate ostensibly new pieces of culture in a matter of seconds. AI art is already replacing that of living artists and writers for album covers, journalism, and other content creation online.

However, AI generators cannot generate genuinely new artworks or thoughts that respond to either our historical and cultural epoch nor our psychological and phenomenological life as human beings. AIs like Midjourney and ChatGPT work through combination, ‘trained’ through the input of large amounts of data to combine already-existing work in response to a user prompt.

As such, the alterity, the otherness that can be found in every genuinely creative piece of art is unlikely to be found in AI-generated work. AI cannot, in the famous words of the modernist poet Ezra Pound, ‘make it new,’ invent new forms, new aesthetics, new ideas. While the pictures from or Midjourney or Dall-E are amusing enough, they exist largely in the space of the mash-up–one familiar reference point matched with another.

As the cultural critic Brandy Jensen said on Twitter apropos of ChatGPT, ‘are we supposed to think this is good?every post i’ve seen is people going ‘oh my god…. jesus fuck….wow’ to absolute dogshit writing.

Where the deconstructionist Paul de Man once argued that the language of literature has a kind of inhumanity to it—wherein we feel that language is speaking us and not vice versa—the allure of AI-generated art is a fetishism of technology itself: AI speaking us, and not vice versa.

The standardisation of feeling that Jameson noted is already leading to an intensified standardisation of aesthetic norms, of ideas, of content becoming ever more banal across platforms. This is an intensified form of what the British writer Mark Fisher termed ‘the slow cancellation of the future’—the diminished capacity of art in the neoliberal era to move us into new politicised relationships in response to the shock of new aesthetic forms. As Fisher so trenchantly reminded us, art can and indeed must open up new pathways beyond the already-known.


Just as importantly, artists and writers might lose the ability to support themselves in order to develop their skills enough to make something genuinely new. AI art will undoubtedly take the jobs of people making online content, musicians making backing music, visual artists designing marketing materials, writers, journalists. Sure, most of that is bad art, if you can call it that at all. But it also hones the skills and pays the bills of artists so they might at some point make something worthwhile. Already, artists have protested their work being used to train AIs without payment for this usage, let alone the content derived from their distinctive style.

AI generators are an appropriation of the commons of artistic creation by Silicon Valley. Not only is that wrong in itself, but in the long run it is likely to produce an ever-decreasing pool of new content to drive work-generation. As such, the advent of AI suggests something profoundly troubling: a further diminution of our capacity to make art, and think.

In its combination of the already-known, AI cannot respond to the important challenges of our age–climate change, various forms of prejudice, the inequalities of capitalism and settler colonialism, incipient fascisms. We need, more than ever, art and thought that gives us the authentically new, that tells us something about human life beyond the norms of media normativity of tech platforms and billionaire tyrants.

AI cannot ‘think’ the historical situation we are in, nor can it respond in ways beyond the limits of normativity. Produced through the rigidity of algorithms, an AI cannot decide when a law is unjust, when we need to take unlawful action and recalibrate the political relationship between the government and the governed for the greater good. Automation cannot respond ethically to the Other in thought or art alike–in fact, it is more likely to reproduce dominant forms of prejudice, as in the transphobic AI-generated Seinfeld episode or chatbots that descend into racism.


What I suggest, therefore, is a return to a qualified humanism.

Artists need to embrace the aura of the singular piece of artwork–the site-specific work, the small-run pressing, the live performance. In other words: the material. AIs cannot reproduce a material, one-of-a-kind object, or a time-and-space specific experience. Given the conditions under which artists create in our present age, the turn to the material is inextricably political—not merely a capitulation to the fetishism of the rarified collector but a resistance to the large-scale appropriation of mechanical reproduction.

In his book Scorched Earth, the art theorist Jonathan Crary has recently talked about the ‘impossibility of life-affirming and non-financialized uses of the internet,’ the foreclosing of genuine forms of political collectivity possible when people gather together. While early artistic and activist response to the internet saw it as a space for emancipatory discourse, the ever-intensifying enclosure, quantification and commodification of basically every aspect of online life has extinguished much of this potentiality.

It is not that we cannot make use of technology in making art or thinking, but rather that the business model of platform capitalism is primarily making use of us—its goal is capital accumulation, not human (let alone non-human) flourishing.

Posthuman theorists of all political stripes—from feminist to POC to disabled to queer—have noted the ways that the category of the human has excluded certain others throughout modernity, as well as the agency of non-human others like animals and plants. We cannot simply return to a politically retrogressive model of the human that is normatively white, male and heterosexual, nor continue to ignore our impact on the environment around us. Without heading into reactionary nostalgia, it is time for an artistic and intellectual turn to embodied relationality to one another and the non-human—to a joyful affirmation of what it means to experience and feel unmediated by the algorithmic.

And we as an audience need to resuscitate the author in turn, to care not only about the product but the life of the person or people who have made it. AIs cannot make something that thinks the world in a new way, that asks us to reconsider the familiar from a new perspective—whether it be a historically-marginalised minority, a political position that has been shut out of media conversation or a virtuosic performance, let alone the aesthetic breakthroughs of form that have characterised every great art movement from modernism to punk. We must demand more from our art than simply the thing in itself, divorced from a context in which a work makes sense. Art is not just ambient, and we lose important parts of ourselves when we think of it as just content and not an expression of a living being.

Personally, I think we should simply ban such generators wholesale in this country, given the minimal social benefit they have and utter devastation to education, the arts, journalism and other sectors they are likely to produce.

The privatisation of much of the history of human knowledge for the benefit of a few Silicon Valley types in exchange for dogshit content seems a bad bargain to me, and it is profoundly troubling that this has not been more central to the cultural conversation.

Or we might, if not ban AI, at the very least nationalise it. If it is to be ‘trained’ on our common artistic and intellectual heritage, why should the rewards and wealth it creates not be similarly collective?  As Crary has said, ‘part of our current crisis is the indifferent acceptance of the now banalized notion that our future is being invented by a small number of powerful corporations.’ The malaise of inaction in which we find ourselves is an indicator of a lack of belief in true democracy–our collective ability to change the conditions under which we work and live.

As Benedict Spinoza once wrote, ‘no one has yet determined what the body can do,’ and we must embrace our fugitive corporeality as we make and consume art and thought grounded in genuine experience in this new age of AI automated content.


Image: Carel Willink, City View (1934)

Emily McAvan

Emily McAvan is an Australian literary critic and theorist.

More by Emily McAvan ›

Overland is a not-for-profit magazine with a proud history of supporting writers, and publishing ideas and voices often excluded from other places.

If you like this piece, or support Overland’s work in general, please subscribe or donate.

Related articles & Essays

Leave a Reply

Your email address will not be published. Required fields are marked *