Authorship as Guarantee
AI destabilizes authorship more thoroughly than any previous technology, because it intervenes not at the level of reproduction or distribution but at the level of generation. The printing press reproduced the author's words. The camera reproduced the world the photographer framed. The recording captured the musician's performance. In each case, the human was the source and the technology was the means of transmission. AI reverses this relationship: the machine generates; the human selects, directs, evaluates — but does not, in the traditional sense, make. Groys's framework suggests three possible resolutions, corresponding to different understandings of what authorship is for.
In the AI Story
The first resolution reconceives the author as guarantor. The author is not the person who made the work but the person who takes responsibility for it — who stands behind its claims, accepts accountability for its errors, and guarantees its quality with her name and reputation. This understanding is already implicit in professional practice: the lawyer who signs the brief is the author regardless of how many associates contributed to it. AI extends this logic without fundamentally altering it. The professional who directs Claude and signs the output is the guarantor. The guarantee is the authorship.
The second resolution replaces authorship with curation. If the machine generates and the human selects, then the human's creative contribution is curatorial rather than authorial. The AI-assisted creator is not a writer who happens to use a tool. She is a curator who selects from the machine's production the outputs that merit attention, frames them within a context that gives them meaning, and presents them to an audience whose reception completes the cultural circuit. This understanding aligns with Groys's analysis of the migration from production to curation as the defining shift of contemporary culture.
The third resolution — and the one Groys would find most interesting — embraces indeterminacy as productive condition. When it is genuinely impossible to determine whether a work was produced by a human, a machine, or a collaboration, the question of authorship becomes irrelevant to the evaluation of the work. The work is evaluated on its own terms: by its quality, its capacity to transform the context in which it appears, its relationship to the archive against which novelty is measured. This resolution would represent a genuine advance in cultural logic — a liberation from the biographical fetishism that has distorted aesthetic judgment since the Romantic period.
The submedial space provides a final framework for understanding what authorship means when the fiction of total control has been abandoned. Every text has a visible surface and a hidden depth. The author of the AI-assisted text may not have generated every sentence on the surface. But she holds the depth. The experiences the text describes are hers. The judgment that selected and shaped the output is hers. The biographical position from which the text speaks — the specific angle of vision that only this person, with this history, in this moment, could produce — is hers. Authorship, reconceived through Groys's framework, is not surface control. It is depth.
Origin
The concept of authorship as guarantee draws on Groys's longer engagement with the institutional constitution of cultural authority, developed across Art Power (2008) and Going Public (2010). The framework was applied to AI directly in the 2023 e-flux essay on prompting, which argued that the writer's body — pressed against the keyboard, laboring at the manual activity of composition — has been the last site of visibly embodied cultural production, and that AI eliminates this body while leaving the structural function of authorship unchanged.
Key Ideas
The guarantor replaces the maker. The author is the one who takes responsibility for the work, not the one who physically produces it — a logic already operative in professional practice and now universalized by AI.
Curation is creation. The selection, framing, and contextualizing of machine-generated outputs constitute a creative contribution distinct from production.
Indeterminacy can be productive. Embracing the impossibility of attributing specific contributions allows evaluation to focus on the work rather than on biographical fetishism.
Depth is the human contribution. The submedial space of experience, judgment, and biographical specificity is what the human brings — and what the machine cannot provide.
Debates & Critiques
The legal and ethical implications of authorship-as-guarantee remain unsettled. If the author is the guarantor rather than the maker, what happens when the guarantor has not fully examined the material she guarantees? Groys's reply is that this condition is not new — authors have always been guarantors of material they have not independently verified — but AI makes the structural fact unignorable.
Appears in the Orange Pill Cycle
Further reading
- Roland Barthes, "The Death of the Author" (1967).
- Michel Foucault, "What Is an Author?" (1969).
- Boris Groys, From Writing to Prompting, e-flux Journal #134 (2023).
- Rosalind Krauss, The Originality of the Avant-Garde and Other Modernist Myths (MIT Press, 1985).