Research Question

How does generative artificial intelligence, understood as a co-creative agent, redefine the poetics and methodologies of contemporary installation art, and what implications emerge from this artist-machine relationship? This research investigates the integration of AI-based agentic systems into creative processes, examining their operational modes, meanings, and epistemological status. In doing so, it reactivates a reflection on surface, image, and gesture as spatially situated elements, with a view to a possible reformulation of the visual arts within an installation-based and computational framework.

Methodology Framework

The research is based on a critical selection of academic literature, artists’ statements, and interviews with curators and artists. It adopts a practice-based approach, combining qualitative methods with installation-based experimentation. The theoretical framework is structured around two complementary perspectives: the historical, focused on the evolution of installation art in relation to computational media as procedure, environment, and system, as well as on the transformations that have redefined its status from the 1950s to the present day; and a phenomenological perspective, concerned with the transformations of the creative process within agentic ecosystems. From this perspective, the works of Grégory Chatonsky, Hito Steyerl, and Pierre Huyghe are examined in relation to generative processes and emerging forms of transmedial imagination. The international research component is developed in two stages: at Sorbonne Nouvelle University, under the supervision of Professor Antonio Somaini, where the semantic organisational processes performed by neural networks within latent spaces have been investigated; and at FACT Liverpool, where fine-tuning methodologies for LLM-based systems will be further explored in support of the development of the proposed installation project.

Case Studies

The case studies function as situated experimental devices for investigating human-AI co-creative dynamics. The project involves the creation of a ‘confessional’ structure composed of industrial metal sheets, conceived as a relational installation within a productive environment and aimed at collecting workers’ testimonies through conversational AI systems. The collected materials will subsequently be reinterpreted through collaboration with a generative agent and translated into sculptural forms and micro-installations through a dialogical process between human intuition and algorithmic logic. Installation practice thus becomes an operational tool for testing the theoretical hypotheses emerging from the analysis of recent artistic poetics associated with generative processes.

Data Analysis and Interpretation of Results

The qualitative analysis of the collected data aims to highlight the co-creative dynamics emerging between human participants and artificial agents. The expected outcomes include the development of a conceptual framework for human-AI co-creation; a contribution to current debates on authorship and distributed agency; the creation of a poetic archive of manual labour derived from dialogues between construction workers and a conversational chatbot, collected in either digital or printed form; and the production of installation prototypes based on generative methodologies. Selected Bibliography Ackerman, A., Gefen, A., Somaini, A., & Viewing, P. (2025). Le monde selon l’IA: Explorer les espaces latents. JBE Books / Jeu de Paume. D’Isa, F. (2023). La rivoluzione algoritmica delle immagini. Luca Sossella.