The Dominion of COPY

To whom does the sentence you are reading belong?

And the track you have just listened to?

Can we really share that image we stumbled upon while scrolling?

Copyright is the air we breathe: words, sounds, code, toys, even certain industrial shades of color. Modernity has clothed creativity in a legal armor that protects, selects, and at times excludes. Born in eighteenth-century England as a barrier against pirated editions and as a safeguard for writers, copyright has, over the centuries, exceeded its original purpose. It has become infrastructure and business model, a machine of revenue extraction. We now inhabit a paradox in which culture longs to be a commons yet remains fenced in by rights that are necessary, but also hypertrophic.

Two recent books offer a way to measure this fault line. The Capitalism of Creativity by David Bellos and Alexandre Montagu reconstructs the genealogy and distortions of copyright, while Critique of ChatGPT by Antonio Santangelo, Alberto Sissa, and Maurizio Borghi interrogates the blind spot opened by generative artificial intelligence. They are complementary readings: the first traces the architecture of copyright, the second exposes its fissures at the frontier of our present.

Long before art became a negotiable asset, authorship was primarily an ethical notion, a matter of custom rather than of law. When Plato raged against his disciple Hermodorus, who had disseminated lecture notes not intended for publication, he did not accuse him of theft but of violating a tacit code of intellectual decency. In the ancient theater, accusations of plagiarism were matters of professional conduct rather than courtroom disputes. The noble contest with the masters of the past – Virgil rewriting Homer – defined a continuum of deliberate reappropriations, in which debt did not diminish but rather qualified.

Modernity disrupted this understanding. With the ruling of Donaldson v. Beckett in 1774, the principle of copyright was consolidated as a temporary and public right intended to prevent perpetual monopolies and to ensure circulation. Two and a half centuries later, the same framework – extended, stretched, and bent through contracts and exceptions – underpins conglomerates for whom copyright is both financial lever and barrier to entry. What began as an emancipatory promise has turned into an architecture of artificial scarcity.

The art world, a laboratory for appropriation and difference, has been the most visible stage for these battles. The lawsuit against Jeff Koons over his String of Puppies, which directly replicated a photograph by Art Rogers, showed that a mere change of medium was not enough to guarantee transformation. More recently, the case opposing the Warhol Foundation and photographer Lynn Goldsmith shook the assumption that Warhol’s silkscreens of Prince were inherently transformative, with the Supreme Court ruling that in the context of a magazine license they were not. The outcome suggested that transformation is not an ontological state but a function of use, context, and substitution in the marketplace. Even the Bridgeman case in 1999, which ruled that exact photographic reproductions of works in the public domain lack originality and thus cannot be copyrighted, did little to alter a paradoxical reality: high-quality images of canonical works such as the Mona Lisa are still fenced behind licenses sold by agencies that act as if they own what is no longer ownable. These conflicts are not mere legalities; they are aesthetic verdicts. They demand that art render its difference visible and demonstrate why a copy is not a copy.

Santangelo, Sissa, and Borghi illuminate the absurdities that arise when idea and expression fuse. The so-called merger doctrine observes that there is no meaningful alternative way to write E = mc² without betraying its meaning. In software, in notation, in standardized forms, creativity often resides not in deviation but in compliance. Copyright, in these cases, risks privatizing the inevitable. The result is a landscape of gray zones where public domain images are trapped within contractual restrictions, museum archives are locked behind paywalls that mimic expired rights, and photographs that add nothing still claim everything.

Generative AI enters precisely here, amplifying and complicating the terrain. These models absorb texts, images, and sounds, not to understand them but to predict patterns, to produce the most probable output. The legal question is not only what they generate, but from what they generate. The datasets that sustain them contain copyrighted works, public domain material, and reproductions ambiguously claimed as proprietary. Is it lawful to train on such collections? In what measure? Who receives the rent from the output? The European framework, with Directive 2019/790, distinguishes between text and data mining for research purposes, allowed by default, and other uses, where rightsholders can exercise opt-out. Civic tools like Spawning AI’s Have I Been Trained? offer authors a first glimpse into whether and where their works have been absorbed. It is little, but it begins to shift the balance of power. Meanwhile, the major platforms defend these practices in the name of innovation, while many artists experience them as industrial-scale expropriation of cultural memory. This is not a dialogue of the deaf but a conflict over who decides the conditions of transformation.

Some, like Yanis Varoufakis, have spoken of techno-feudalism: a world where those who control infrastructures, datasets, and distribution channels also control the means of symbolic production. Museums become training data providers, galleries turn into licensing nodes, archives mutate into gated repositories. Conservation risks collapsing into monetization; curation risks being reduced to a business model.

The challenge is not to abolish copyright, nor to embrace technological determinism. What is required is an institutional imagination that balances transparency, opt-out mechanisms, collective licensing, and clear exceptions for transformative use in art. These are not instruments of war but of equilibrium, designed to prevent complexity from becoming an instrument of deterrence.

What do we ask of art today? Not innocence, but consciousness of genealogies and responsibility for difference. Appropriation, when it occurs, must declare itself as visible rewriting, as montage that acknowledges its sources rather than neutralizing them. Technology, if it is to be embraced, must come with the visor lifted, with full disclosure of datasets and methods. Museums, archives, and foundations can lead this maturity by opening catalogues for fair use, experimenting with non-predatory licensing, and educating audiences that sharing is not the same as gratuitousness but an act of responsibility.

Bellos and Montagu remind us that copyright was born to limit private monopolies in the name of the public. Santangelo, Sissa, and Borghi show that AI, far from being an oracle, is a mirror: it reflects our way of administering memory. We must decide whether culture will be a fenced garden or a navigable landscape, where paths are drawn not to forbid but to make access equitable, recognizable, and fertile. There is no creativity without inheritance, and no inheritance without fair rules. The critical and curatorial task of our time is precisely this: to transform the empire of copies into a republic of differences.

© Charlotte Madeleine Castelli | All rights reserved