Actor-Network Theory in the Age of Machine Learning: Redefining Artificial Intelligence as a Non-Human Actor in Architectural Design Education
Actor-Network Theory in the Age of Machine Learning: Redefining Artificial Intelligence as a Non-Human Actor in Architectural Design Education
Hossein Moradi Nasab1 Gholamhosein Naseri2 Kamran Javaheri3 Hana Avani4
1) Department of architecture, Se.C, Islamic Azad University, Semnan,Iran,
2) Department of architecture, Damghan Branch, Islamic Azad University, Damghan,Iran,
3) Department of architecture, Se.C, Islamic Azad University, Semnan,Iran,
4) Department of architecture, Se.C, Islamic Azad University, Semnan, Iran,
IOI Code
:
AAHF-ABGHCD
Publication :
National Conference on Architecture and Art Technologies, Smart City, Soft and Creative Industries and Technological Culture(icact.ir)
Abstract :
Generative artificial intelligence has recently entered architectural design studios, blurring traditional boundaries between tool and agent, human and machine, creativity and computation. However, most previous studies have either considered AI as a passive instrument or focused on quantitative performance optimization, neglecting to analyze its agency within the network of architectural education. This study aims to ontologically redefine AI as a non-human actor in the architectural design education process, seeking to answer three fundamental questions: What kind of agency does AI manifest? How does this agency influence students design decisions? And what conceptual model based on Actor-Network Theory can be proposed to analyze this network?
The present research employed a qualitative-interpretive approach using a sequential digital ethnography design in three phases (contextual mapping, translation tracking, and network feedback). The research context was a graduate-level architectural design studio that explicitly incorporated interaction with generative AI systems into its curriculum. Participants included twelve students and two instructors, whose transactional data were collected over fourteen weeks through screen recording, system logs, field notes, and stimulated recall interviews. Data analysis followed the methodological principles of Actor-Network Theory, focusing on identifying moments of rupture and patterns of translation.
The findings revealed that AI manifests four distinct patterns of translation in practice: alternative proposition, resistance, acceleration, and invisibility. These patterns indicate a distributed, context-dependent agency. Forty-seven moments of rupture were identified in the data—points where the design trajectory suddenly and unexpectedly changed, with this change attributed to AI s action. Comparison of the network map at the beginning and end of the semester revealed a transformation from a student-centered network to a distributed network with fluid agency. Two distinct groups of students were identified: those who developed competence in collaborating with non-human actors and those who consistently sought to restore human control.
The conceptual model of the Intelligent Studio Network was formulated across four layers: transactional infrastructure, translation space, institutional framework, and network discourse. Based on this model, design competence in the age of generative AI is redefined from an individual s independent design ability to the capacity for effective participation in a network of human and non-human actors. The study concludes with recommendations for redesigning the architectural curriculum, including rupture-oriented assignments, network review procedures, and teaching translation literacy to instructors.
Keywords : actor network theory generative artificial intelligence digital ethnography architectural design education intelligent studio non human agency