The expanded field of AI production applies Krauss's 1979 structural method to the categorical crisis AI has produced. The traditional binary organizing creative production—the maker who conceives and executes versus the audience who receives—cannot accommodate work produced through sustained human-AI collaboration. The binary assumes a unified locus of creation; the collaboration distributes creation across human intention, machine generation, and the iterative process connecting them. The expanded field generated from the binaries human-intention/not-human-intention and machine-generation/not-machine-generation reveals at least four distinct positions: (1) human-as-sole-maker (the novelist at her desk, the position the old binary describes); (2) human-as-director (the user who specifies intent and evaluates output, analogous to Sol LeWitt or a film director); (3) human-as-collaborator (the user in iterative conversation with AI, where direction and response co-evolve); (4) AI-as-autonomous-generator (the theoretical limit case current systems approach but do not reach). These positions are not discrete categories but points on a continuum, and most actual AI-assisted practice moves between them. The framework's value is not that it settles the question of attribution but that it makes visible the structural positions the dissolution of the authorship binary has opened.
The traditional creative-production binary assumed that conception and execution were either unified in a single maker or divided through a process of complete specification (the architect's blueprint, the composer's score). AI disrupts this assumption by introducing a mediating intelligence that translates intention into execution through a process that is neither fully specified nor fully autonomous. The AI interprets the user's prompt—filling gaps, resolving ambiguities, making choices the instruction did not specify. The output is shaped by user intention, model training, and the dynamics of their interaction, and no single term determines the result. This triadic structure (intention-translation-execution) is unprecedented in the history of creative production and is what renders the traditional binary inadequate.
Segal's Trivandrum training scene provides empirical grounding. The backend engineer building interfaces did not become a frontend developer—she occupied a new position defined by the intersection of her backend expertise, Claude's frontend execution, and their conversation. The traditional categories (backend/frontend) persisted as descriptions of knowledge domains but ceased to function as descriptions of who could build what. The coordination bottleneck that enforced those boundaries—the years of training required to translate intention across disciplinary gaps—was absorbed by the tool, and the absorption produced not the elimination of disciplines but the expansion of the field in which practitioners operate.
The human-as-collaborator position deserves particular attention because it is the position that most fully exploits the field's expansion while being most difficult to evaluate. In this position, the user does not merely direct but enters a process of mutual refinement—the AI's responses reshape the user's questions, which reshape the AI's responses, in an iterative loop producing output neither party could have generated independently. Segal's laparoscopic surgery insight—the connection between ascending friction and surgical technique that emerged from his conversation with Claude—exemplifies this position. Neither the human nor the AI "authored" the insight; it emerged from the collaboration, from the specific intersection of Segal's question and Claude's associative range.
Evaluating output from the collaborative position requires frameworks the inherited criteria (originality, authenticity, individual skill) cannot provide. The relevant questions are not "who wrote this?" but "what does this configuration reveal about the intersection of human intention and machine capability?" and "is the specificity of that intersection sufficient to justify the output's existence?" These are harder questions than the attribution question, and they demand the exercise of structural awareness—the capacity to see through surfaces to the conditions of production—that the aesthetics of the smooth systematically erodes.
The concept originates in this volume's first chapter as the direct transposition of Krauss's 1979 expanded field of sculpture onto the domain of AI-assisted creative production. The transposition is not metaphorical—the structural logic is identical. Both fields are generated from binary oppositions; both produce positions that were logically possible before they were materially occupiable; both require evaluative frameworks adequate to positions the inherited categories cannot accommodate. The AI application is original to this text but structurally continuous with Krauss's method.
Dissolution of the authorship binary. Human creator / tool user cannot describe positions where creation is distributed across intention, generation, and iterative refinement.
Four positions minimum. Sole-maker, director, collaborator, autonomous-generator—each defined by relationships between intention and generation rather than by content.
Positions exist on a continuum. Most practices move between positions within a single session—the field is not a taxonomy of fixed types but a space of variable configurations.
Collaboration as distinct position. The iterative conversation where responses reshape questions and questions reshape responses—neither direction nor execution but their mutual constitution.
Evaluation requires positional awareness. Judging collaborative output by sole-maker criteria is category error—the adequate framework must address the specificity of the position the work occupies.