Agentic Architecture: Synthesising Complexity for Regenerative Futures

Research Background: This work sits at the intersection of architectural design, AI, and systemic urbanism, addressing the urgent need for design methodologies that respond to planetary-scale complexity. It builds on and critiques traditions in computational design, ecological urbanism, and high-resolution material systems, referencing thinkers such as Luciana Parisi (Contagious Architecture), Mario Carpo (The Second Digital Turn), who identifies this work as emblematic of the second digital turn—a paradigm defined by algorithmic authorship, mass-customisation, and non-standard material logics. Carpo positions it as a key exemplar of the shift from parametric formalism to computational design ecologies that embed information directly into material and tectonic systems and Benjamin Bratton (The Stack), and Albert-László Barabási (network science). The project extends ongoing discourse around AI-augmented design cognition beyond automation or optimisation, toward generative systems aware of material, ecological, and societal entanglements. Research Contribution: Agentic Architecture: Synthesising Complexity for Regenerative Futures is a multi-channel video installation exhibited at the 2025 Venice Architecture Biennale (curated by Carlo Ratti). The work presents an original AI-driven tectonic system—AI Timber—combining ancient dry-joint logic with AI combinatorics, enabling adaptive, prefabricated, high-density urban infrastructures. It introduces a novel workflow linking n-dimensional voxelised data, multimodal generative AI, and generative algorithmic systems, including simulations of electromagnetic fields, site-adaptive Multi-Agent Systems, and field-based logics. While not explicitly robotic in this iteration, the work establishes a computational substrate that naturally connects to robotic assembly in future developments. The project consolidates two decades of research in discrete design, simulation, and systemic urban strategies into a speculative but technically grounded prototype for planetary urbanism. Research Significance: This work was directly invited by the curator to participate in the main international exhibition at the Venice Biennale—one of the most prestigious platforms in architecture—positioning it among the top global contributions in the field. Supported by RMIT, AIARCH, and RACE HPC, the work synthesises a suite of methods and ideas developed over years of research, which have also informed prior keynote engagements at forums such as CogX, TU Munich, and the Ecocity Summit. This work has attracted interest from global design-tech actors and urban innovation networks, highlighting its potential for wider impact and translational research collaborations. Its strategic and aesthetic originality, combined with its methodological depth, exemplifies high-impact design research and reflects RMIT’s global leadership in AI-driven architecture.研究背景：本研究立足于建筑设计、人工智能（AI）与系统城市主义的交叉领域，旨在回应应对行星尺度复杂性的设计方法的迫切需求。本研究既继承又批判了计算设计、生态城市主义以及高分辨率材料系统领域的传统研究脉络，参考了卢西亚娜·帕里西（Luciana Parisi，著有《传染性建筑》）、马里奥·卡尔波（Mario Carpo，著有《第二次数字转向》）等学者的理论——卡尔波将本研究视为第二次数字转向的标志性成果，该范式以算法创作、大规模定制与非标准材料逻辑为核心特征。卡尔波同时将其视作从参数化形式主义转向计算设计生态的关键范例，这种计算设计生态将信息直接嵌入材料与建构系统（tectonic system）；此外本研究还借鉴了本杰明·布拉顿（Benjamin Bratton，著有《堆栈》）与阿尔伯特-拉斯洛·巴拉巴西（Albert-László Barabási，网络科学领域学者）的研究。本项目将当前围绕AI增强设计认知的讨论从自动化或优化范畴拓展至能够感知材料、生态与社会关联的生成式系统研究方向。 研究贡献：《智能体建筑：为再生未来整合复杂性》（Agentic Architecture: Synthesising Complexity for Regenerative Futures）是一件多通道影像装置作品，于2025年威尼斯建筑双年展（由卡洛·拉蒂策展）展出。本作品提出了一套原创的AI驱动建构系统——AI木材（AI Timber），将古老的干连接工艺（dry-joint logic）与AI组合算法相结合，可实现自适应、预制化的高密度城市基础设施。本研究引入了一套全新的工作流，将n维体素化数据（voxelised data）、多模态生成式AI与生成式算法系统（包括电磁场仿真、场地自适应多智能体系统（Multi-Agent Systems）以及场域逻辑）进行联动。尽管本版本未明确集成机器人技术，但本作品构建了一套计算基底，可在未来开发中自然衔接机器人装配流程。本项目将二十年来在离散设计、仿真与系统城市策略领域的研究成果整合为一套面向行星城市主义的思辨性但具备技术落地性的原型方案。 研究意义：本作品由策展人直接邀请参展威尼斯建筑双年展主国际展——该展览是建筑领域最具影响力的平台之一，这一参展资格将本研究置于全球建筑领域顶尖成果之列。本研究获得了皇家墨尔本理工大学（RMIT）、AIARCH以及RACE高性能计算集群（RACE HPC）的支持，整合了多年研究中形成的一系列方法与理念，这些成果也曾为CogX、慕尼黑工业大学（TU Munich）以及生态城市峰会等论坛的主旨演讲提供了理论支撑。本作品已吸引全球设计科技机构与城市创新网络的关注，彰显了其在更广泛领域产生影响以及开展转化研究合作的潜力。其战略与美学层面的独创性，结合深厚的方法论底蕴，彰显了高影响力设计研究的特质，同时也体现了皇家墨尔本理工大学（RMIT）在AI驱动建筑领域的全球领先地位。