An Ecocentric Assessment Framework for Artificial Intelligence and Computational Design in Architecture

Artificial intelligence (AI) supports environmental performance prediction and the evaluation of alternatives in architectural design. However, improvements in energy or carbon indicators alone do not demonstrate that non-human life holds independent value in design decisions. This chapter examines AI and related computational design approaches in architecture from an ecocentric perspective through a focused narrative review and conceptual synthesis supported by a structured database search. The 375 raw records retrieved from Scopus and Web of Science were reduced to 256 unique records; following title and abstract screening, a final purposive selection of 19 studies was established. Together with one supplementary publication, the 20 documents were examined under the themes of performance and decision-making, life cycle boundaries, the representation of non-human life, habitat geometry and materials, time and maintenance, and visual generation and ecological validation. The synthesis informs a preliminary ecocentric assessment framework comprising three dimensions: stakeholder scope, integration of environmental and ecological impacts into decision-making, and ecological responsibility. By distinguishing computational success from ecological outcomes, the framework provides an analytical basis for examining who benefits from design decisions, how environmental and ecological requirements influence those decisions, and how responsibility for their outcomes is addressed.