Appropriation is Not New: The Eternal Return of the Vanguard

Here’s the thing about originality, vanguard and how things get absorbed into the mainstream. Spoiler: it’s not new, just faster now with AI.

We need to understand how novelty works in culture, so that we calm down about AI homogenization. The anxiety of derivative work and the role of the original artist creativity have not changed. Here are a few ideas to consistently think about this, and the roles that may pay off today.

---

It’s a cultural cycle of new, copy, paste and new

Eco's key insight, developed across Opera Aperta (1962) and Apocalittici e Integrati (1964), is that the avant-garde and mass culture exist in a productive tension rather than a simple opposition. The avant-garde produces formal novelty — a new way of organising a text, an image, a sound. Mass media, by its nature, captures that novelty, strips it of its original disruptive intent, and repackages it as consumable sensation. Eco's definition of kitsch is precise on this point: it is the prefabrication and imposition of effect — borrowed formal moves deployed not to challenge the audience but to produce a predictable emotional response. The process is not theft in any legal sense; it is absorption. The avant-garde gesture gets flattened into a style, the style becomes a convention, the convention becomes wallpaper. And because the original gesture has been emptied of its destabilising power, a vacuum forms — a space where the conventions feel exhausted and something genuinely new becomes necessary again. This is the cycle: disruption, absorption, exhaustion, disruption.

Vanguard kicks the door open, the mainstream absorbs it, and the cycle restarts

What matters for your argument is that this cycle has always depended on derivative work. The absorption phase (this long middle of the cycle where the original idea is diluted, recombined, adapted, applied to contexts its creator never intended) is not a failure of culture. It is the mechanism by which formal novelty becomes legible to a broader audience, which in turn creates the conditions for the next disruption. The Bauhaus grid entered corporate branding. Punk typography entered magazine design. Brutalist web aesthetics entered fintech landing pages. At each stage, something was lost — the ideological charge, the roughness, the refusal — but something was also gained: a wider repertoire of formal possibilities available to anyone who needed to make something. The derivative work is the distribution system for originality.

AI mechanises the derivative phase. That is all it does, structurally speaking, and that is why the cycle is not broken, only accelerated. Things get older faster. Is it good? Is it bad? Do we need to think in these terms, even?

When a generative model absorbs a novel visual grammar and begins producing competent variations of it within weeks, it is performing the same operation that mass media has always performed: capturing formal novelty and repackaging it for consumption. The difference is speed and volume. A design pattern that might have taken five years to move from avant-garde publication to mainstream application now makes that journey in months, sometimes less. The absorption phase compresses.

This compression has two consequences that pull in different directions. The first, which the homogenization literature documents well, is that the exhaustion phase arrives faster. When AI tools recombine the same patterns at scale, the feeling of staleness sets in sooner — the visual equivalent of hearing a song played too many times. Conventions that once felt fresh for a decade may now feel depleted within a year or two. The second consequence, less discussed, is that the vacuum phase also arrives faster. If exhaustion creates the demand for novelty, and if exhaustion is now accelerating, then the system's appetite for genuinely original ideas is increasing, not decreasing. The cycle spins faster, but it still depends on the same fuel: someone, somewhere, producing a formal gesture that has not been made before.

Eco was careful not to take sides in the apocalyptic–integrated debate. He was suspicious of intellectuals who condemned mass culture wholesale (the apocalyptics), because their position implied that the masses were incapable of critical engagement. He was equally suspicious of those who embraced mass culture uncritically (the integrated), because that position ignored the real mechanisms of manipulation and dilution at work. His position was analytical: understand the cycle, understand the forces, understand what is gained and what is lost at each stage. Applied to the AI moment, this means resisting both the panic that AI will destroy originality and the optimism that AI will democratise it. The more precise claim is that AI accelerates the derivative phase of a cycle that has always existed, and in doing so it increases the structural demand for the kind of original work that the derivative phase feeds on.

Most of us are between newness and mass market

For professionals in design, strategy, and product (people who operate in the territory between the avant-garde and the mass market) the implication is that the derivative work they do with AI is not inherently less valuable than the original work that preceded it. Adaptation, translation, application to new contexts: these are legitimate and necessary acts within the cycle. But the implication is also that someone must continue to produce the original gestures that keep the cycle turning.

If the absorption phase is now mechanised and the exhaustion phase is accelerating, the bottleneck moves upstream, to the moment of genuine novelty. The people who can produce that novelty occupy a structural position in the creative economy that AI has made more important, not less, precisely because the machinery that distributes their ideas now operates at a speed that would have been unimaginable to Eco's generation.

How very Benjaminian of you

Walter Benjamin diagnosed a version of this dynamic in 1935, before mass media had fully matured. In The Work of Art in the Age of Mechanical Reproduction, he argued that the original artwork possesses an aura — a quality of unique presence tied to its existence in a particular place and time, embedded in the tradition and ritual context from which it emerged. Mechanical reproduction — the lithograph, the photograph, the film print — detaches the work from this context and makes it available to audiences who would never encounter the original. The aura decays in the process. A poster of a Vermeer in a student flat is not a Vermeer; it is a Vermeer-shaped object that has been stripped of the specific qualities that made it singular. Benjamin was ambivalent about this loss rather than simply mournful. He recognised that reproduction democratised access to art, liberated it from what he called the parasitical dependence on ritual, and opened it to new forms of political engagement. The masses could now encounter works that had previously been available only to those who could travel to the right church or the right gallery. Something was lost — the aura, the authority of the original — but something was also gained: a new kind of cultural participation that was collective rather than contemplative.

Benjamin's framework maps onto the avant-garde absorption cycle with striking precision, but it also requires an important update when applied to AI. Mechanical reproduction, as Benjamin understood it, copies the work — the specific painting, the specific photograph, the specific film. The aura of the individual object is lost because identical copies circulate in its place. AI reproduction operates at a different level: it copies the style, not the work. When a generative model absorbs a designer's visual language and begins producing competent variations, it is not reproducing any single artefact — no specific poster, no specific layout, no specific illustration is being duplicated. Instead, it is reproducing the formal grammar that made those artefacts recognisable as the work of a particular sensibility. The aura that decays is not the aura of the object but the aura of the idiom. A design approach that once felt distinctive because it was associated with a specific person, studio, or movement becomes a reproducible style — available to anyone with access to the right model and the right prompt, detached from the biography and the context that gave it meaning.

This is a deeper form of the loss Benjamin described, because it operates upstream. Mechanical reproduction left the style intact; it merely multiplied the objects. A thousand lithographs of a Toulouse-Lautrec poster do not diminish the recognisability of Toulouse-Lautrec's visual language — if anything, they reinforce it, because the style becomes associated with the reproduced image. AI reproduction dissolves the style itself into a palette of recombineable elements. The visual grammar that once belonged to a particular practice becomes available as a parameter, adjustable and detachable from its origin. In Benjamin's terms, what is lost is not the aura of the work but the aura of the making — the sense that a particular way of seeing was anchored in a particular life and a particular tradition of practice.

At the same time, Benjamin's ambivalence remains instructive. If the democratisation of the work had political and cultural value — if bringing art to the masses opened new possibilities for engagement — then the democratisation of the style may carry a similar ambivalence. When a visual grammar becomes reproducible, it becomes usable by people who could not have invented it but who can now deploy it in contexts its originator never anticipated. A design language developed in a Dutch studio for cultural institutions might, through AI-mediated reproduction, find its way into a health-tech startup in Tallinn or a civic engagement platform in Lisbon. The originator's aura is diminished, but the formal vocabulary enters wider circulation, and the cycle continues: exhaustion of the reproduced style creates demand for the next genuinely original one.

The radical solution: originality?

While RPM has changed dramatically, the cycle isn’t new. Which may create an interesting detour: instead of perfecting a carbon-copy of some new trend (replicating design trends on a website, brand or t-shirt line), creatives may realize that the effort of copying won’t pay off. The machine will exhaust everyone very quickly with that same style. So a better avenue is to, well, think of something new and be the one being copied instead.

---

Sources to cite:

• Benjamin, W. (1935/2008), The Work of Art in the Age of Mechanical Reproduction, trans. J.A. Underwood, Penguin. (Originally published as "Das Kunstwerk im Zeitalter seiner technischen Reproduzierbarkeit," in Zeitschrift für Sozialforschung, 1936.)
• Eco, U. (1962/1989), Opera Aperta / The Open Work, Harvard University Press. Especially Chapter 9, "The Structure of Bad Taste," where Eco defines kitsch as the prefabrication and imposition of effect.
• Eco, U. (1964), Apocalittici e Integrati, Bompiani. The title essay on the dialectic between mass culture critics and mass culture enthusiasts.
• Doshi, A.R. and Hauser, O.P. (2024) for the empirical evidence of homogenization that maps onto the absorption/exhaustion dynamic.