From generative algorithms to Renaissance workshops: a parallel in mechanisms and reconfiguration of artistic production systems.

The rapid development of generative artificial intelligence (AI) is transforming the conditions under which visual images are conceived, produced, and valued. This thesis examines generative AI not as an autonomous creative agent, but as a distributed system of artistic production involving models, data, developers, artists, users, and platforms. Drawing on Howard S. Becker’s theory of Art Worlds (1982), it investigates how such distributed contributions are transformed into outputs that can be attributed, controlled, and economically appropriated. The thesis develops the concept of institutional closure to describe the mechanisms through which collective creative activity is consolidated within an organizational structure. It identifies two distinct forms: Dynastic Closure, characteristic of the Renaissance family workshop, and Platform Closure, characteristic of contemporary generative AI infrastructures. The comparison is structured around three interconnected dimensions—authorship, ownership, and remuneration—and combines organizational theory, art history, technology, economics, and legal analysis. The historical case study focuses on the workshop of Jacopo Bassano (c. 1510–1592), examining its division of labour, serial production, modular reuse, and family organization as a form of collective artistic production. The analysis compares this system with contemporary generative AI architectures, particularly Generative Adversarial Networks (GANs) and Creative Adversarial Networks (CANs), identifying a structural analogy in their processes of recombination and variation while avoiding any technological or historical equivalence. The thesis argues that the fundamental difference between the two systems lies in their mechanisms of closure. Whereas the Bassano workshop internalized labour, knowledge, and reputational assets within a hereditary organization, contemporary platforms coordinate distributed contributions through technological and contractual infrastructures while incorporating creative data into proprietary systems. This distinction generates different configurations of authorship, ownership, and remuneration and helps explain the attributional and economic tensions surrounding generative AI. Ultimately, the thesis argues that generative AI does not represent the disappearance of artistic production, but a reconfiguration of distributed creativity, in which value increasingly shifts from physical execution toward the definition of constraints, curation, and critical selection. The comparison with the Renaissance workshop provides a historical framework for understanding these transformations and for reconsidering how collective creativity can be governed and economically organized.