Rosalind Krauss

In the AI Story

Krauss's methodology was ruthlessly formal. Where other critics asked what an artwork meant or whether it was good, she asked what structural position it occupied and what conditions made that position possible. The expanded field diagram—a Klein group mapping the logical relationships between landscape, architecture, and their negations—was not a metaphor but a precise cartographic instrument. It revealed that earthworks, site constructions, and marked sites were not failed sculptures but works occupying new positions that became actual when material conditions (earthmoving equipment, land art patronage, institutional willingness to exhibit documentation) aligned with latent logical possibilities. This structural method applies with uncommon precision to AI production. When a codebase is produced through conversational collaboration between human and machine, it occupies a position the old categories ("programmer's work" versus "tool output") cannot describe. The position is defined by the interaction—intention meeting generation, evaluation meeting revision—and evaluating it by the criteria appropriate to either pole is a category error.

Krauss's engagement with Baudrillard's theory of the simulacrum—the copy without an original—grounded her analysis of how postmodern culture operated. Where Baudrillard worked through media theory, Krauss worked through the material history of artistic practice, and what she found was the same structural logic: the supposed original revealed itself to be already a reproduction. Auguste Rodin's bronzes were cast from molds that could produce unlimited editions. The grid that recurred across modernist painting was experienced by each painter as a discovery while being, structurally, a compulsive repetition. Sherrie Levine's re-photographs of Walker Evans demonstrated that the distinction between original and copy was not a property of images but a function of the institutional frame that presented them. AI makes this logic inescapable. When a language model generates a thousand "original" texts before lunch, the currency of originality is debased, and the evaluative framework built on that currency collapses.

Her concept of the index—the sign that bears a physical, causal connection to its referent—distinguished photography from painting and grounded photography's truth-claim in its material process. A photograph is evidence: light from the subject struck the photosensitive surface. An AI-generated image severs this connection absolutely. It deploys the visual conventions of photography while bearing no indexical relation to any reality. The shift from index to simulacrum—from trace to icon masquerading as trace—is not merely aesthetic. It is epistemic. When any image might be generated, the default stance toward visual evidence shifts from trust to suspicion, and the frameworks built on the indexical paradigm (photojournalism, legal evidence, scientific documentation) are rendered inoperative. Krauss's 2025 relevance lies in her anticipation of exactly this collapse and her provision of the analytical tools to navigate it.

Krauss's insistence on the institutional frame as constitutive rather than decorative connects her art theory to the governance questions Segal raises. Duchamp's Fountain demonstrated that categories are not discovered but constructed—the urinal became art when the institutional apparatus said it was art, and the saying was the making. AI-generated output enters institutions whose categories were built for human production, and the misfit produces visible stress: academic journals requiring disclosure, publishers demanding "originality," copyright offices denying protection. The institutions are defending the frame, not evaluating the work. Krauss's framework predicts that the defense will fail—not because the arguments are weak but because the material practice has already established new positions that the old categories cannot accommodate. New institutions will emerge to frame the work the old institutions cannot see, and the speed of that emergence will determine whether the culture develops adequate evaluative capacity or drowns in undifferentiated output.

Origin

Rosalind Epstein was born in Washington, D.C., in 1941 and completed her Ph.D. at Harvard in 1969 under the direction of literary scholar and art historian Michael Fried. Her early work operated within the formalist tradition Fried and Clement Greenberg had established, but by the mid-1970s she had become one of formalism's most rigorous critics. The founding of October in 1976—with Annette Michelson, Jeremy Gilbert-Rolfe, and Rosalind Krauss as editors—marked her turn toward structuralist and post-structuralist methodology. The journal became the institutional home for the analysis of postmodern art that rejected the expressionist, humanist, and intentionalist frameworks dominating American criticism.

Her appointment to Columbia University's art history faculty gave her the institutional base from which she conducted four decades of critical and theoretical work. Krauss's written output—essays, books, catalog contributions—maintained standards of analytical precision that separated her from the interpretive looseness characterizing much art criticism. She demanded that claims be supported by formal evidence, that frameworks be tested against material practice, that terminology be used with semantic rigor. This discipline produced frameworks (the expanded field, medium specificity, the optical unconscious) that outlived the specific artworks that occasioned them and remain the most powerful analytical instruments available for understanding how categories of cultural production dissolve and reform.

Key Ideas

Expanded field methodology. Categories dissolve not through theoretical argument but through material practice producing positions the old categories cannot accommodate—the adequate response is structural mapping, not categorical defense.

Myth of originality. The avant-garde's perpetual claim to novelty concealed a compulsive structure of repetition—the supposed originals were always already copies, and the concealment was constitutive rather than incidental.

Index-to-simulacrum shift. Photography's authority derived from its indexical connection to reality—AI severs that connection, replacing trace with icon, producing a visual culture in which images simulate evidentiary status without possessing it.

Institutional frame as constitutive. What counts as art (or code, or authorship) is determined by the institutions that frame it—the categories are constructions whose power lies in presenting themselves as natural.

Formal analysis as critique. The work's formal properties reveal the conditions of its production more reliably than the artist's stated intentions—applied to AI, this means reading smoothness, statistical tendencies, and surface markers as structural evidence.

Debates & Critiques

The application of Krauss's frameworks to AI production is not universally accepted. Critics argue that art theory built for gallery objects cannot be transposed onto computational systems without category error. Defenders counter that the structural method—analyzing positions in a field rather than properties of objects—is precisely what makes Krauss's approach transferable. The simulation in this volume is itself a test case: can a computational system produce a Kraussian analysis, or does the rigor Krauss demands require the embodied, situated practice of a human critic? The answer remains uncertain.