From Tool to Meta-Partner: Reframing Architectural Ethics in AI-Augmented AEC Practice

Parallel session

Parallel session 5

Presentation slot

Track

Track 1: Conceptual Disruption: Theoretical Issues and Practical Approaches

Presenter

Dr. Ayça Ayaz Erdağ is an architect, researcher, and educator working in architecture, building science, and digital design. She holds a bachelor’s degree in architecture from Eskişehir Osmangazi University, a master’s degree from the New Jersey Institute of Technology, and a PhD in Building Science from Middle East Technical University, awarded in 2025. She currently teaches architectural acoustics and building safety as a part-time instructor in Bilkent University’s Department of Interior Architecture and Environmental Design.

Her research focuses on AI-assisted design, digital craftsmanship, and architectural knowledge production. Her doctoral dissertation, “New Digital Craftsmanship ‘DeepSmart Corpus’: Learning from Pavilion Data with an AI-Assisted Model,” examined relationships between design, materials, and fabrication through a dataset of 158 pavilions. It explored how knowledge derived from architectural examples can be interpreted through artificial intelligence and inform design processes.

Her work investigates how digital technologies transform human agency, craft knowledge, and design decisions in architecture. She is particularly interested in responsibility and knowledge representation in AI-assisted design. Connecting academic research, architectural practice, and education, she examines the human and ethical dimensions of technological transformation.

Type of presentation

Single paper

Room

Quantum 1

Abstract

Artificial intelligence is increasingly embedded within architecture, engineering, and construction workflows, shaping how design knowledge is generated, decisions are negotiated, and professional expertise is enacted. Yet ethical debates on AI in the built environment often continue to frame artificial intelligence as a neutral instrument subordinate to human intent. This paper argues that this assumption is no longer sufficient. Instead, AI should be understood as an emerging meta-partner: an epistemic and operational actor participating in collaborative design ecologies alongside architects, engineers, computational systems, and data infrastructures. From the perspective of architectural ethics, this shift destabilizes established understandings of agency, responsibility, authorship, and professional judgment. Architectural practice has traditionally relied on bounded ethical roles assigned to identifiable human actors. However, in AI-mediated environments, design intelligence increasingly emerges through distributed interactions among human expertise, predictive systems, optimization engines, simulation tools, and shared datasets. Drawing on the DeepSmart Corpus, a dataset of 158 experimental pavilion case studies involving computational design, robotic fabrication, and machine learning, this paper examines five ethical transformations in AI-augmented AEC practice: the redistribution of agency between human and computational actors; emerging forms of distributed responsibility; epistemic concerns related to data selection, bias, and design knowledge production; tensions between professional autonomy and algorithmically mediated decision-making; and the disruption of architecture’s conventional understanding of tools, creativity, and authorship. In response, the paper proposes ethical meta-partnership as a theoretical framework for rethinking architectural ethics under conditions of distributed design intelligence. By positioning architecture as a critical site where technological transformation becomes materially, socially, and professionally consequential, the study contributes a design-centered perspective to broader debates in the philosophy and ethics of technology in turbulent times.