The impact of Artificial Intelligence (AI) research advancements is visible in many fields, from medicine to the visual arts. In architecture, Generative AI tools have enabled designers to generate large numbers of compelling images through simple textual prompts, triggering critics to focus the theoretical debate around issues of authorship and creativity of design outputs: the process goes into the background, and investigating the collaboration mechanisms between human architects and machines loses its centrality. This essay aims to restore continuity between the research conducted in AI in design of the 1990s and the current technological developments in the field by illustrating how recent applications of AI models and tools can support designers in areas other than image generation. We discuss how architects can establish and strengthen forms of collaboration with AI to develop and review design proposals, offering an approach that embraces AI as a partner rather than a replacement for human creativity. The rapid progress made by Artificial Intelligence (AI) researchers is transforming a niche field of study into one of the most significant innovations of the century, with applications being developed in all disciplines, from medicine to the visual arts.

Birds and many other living organisms rely on large old trees for their survival. However, these trees are rapidly disappearing from many landscapes. The regrowth of such trees takes hundreds of years, and in many disturbed landscapes, wildlife populations cannot persist without temporary human-made replacement structures. In 2022, Mirra et al. [1] conducted a study on the AI-based generation of 3D models, or visual abstractions, of large old trees. An AI agent analysed a dataset of these trees to identify features that appeal to animals and generated forms that approximated these features. The researchers also evaluated the usefulness of the AI-generated forms as artificial replacements for trees using morphological and cost criteria.

Building on this research, we developed our pavilion entry for the IASS 2024 design competition. Our focus was on creating (1) a design strategy to translate the AI-generated visual abstractions into buildable tensegrities, and (2) a fabrication method that facilitates the transport, assembly, and disassembly of these structures using repurposed and biodegradable materials. We named this prototype “FloaTree”. We have constructed prototype tests in Melbourne, Australia, and will build the full-scale pavilion during the annual IASS symposium in Zurich, Switzerland, in August 2024.

This paper describes a novel approach to structural optimisation based on learning design strategies rather than searching for optimal solutions. In the proposed approach, an AI agent is trained through Reinforcement Learning (RL) to explore a 3D modelling environment and iteratively morph a flat NURBS surface into a doubly-curved shell structure. At each iteration of the 3D modelling process, the agent computes the maximum structural displacement through FEM analysis: it learns to select modelling actions through this feedback and progressively improves the performance of the input surface. Unlike current applications of RL in structural design, where the AI agent generates design options by recombining a predefined set of design variables, our approach aims to create structural forms through the interaction of a designer and an AI agent within a 3D modelling environment. An application illustrates that our agent can interpret a preliminary structural form defined by a designer and iteratively transform such a form to improve its structural performance. The trained agent can hence transform the geometry and improve the structural performance of any open surface that features a square footprint and is defined through a sequence of modelling commands. Preliminary results suggest that this AI agent can be used for the development of more interactive tools for structural design and optimisation.

The design of music venues, such as concert halls and open-air concert stages, requires an integrated approach in which the acoustic response of the space being created is evaluated at every stage of the process to inform its formal development and associated performance. Various software exists to assess acoustic performance: Odeon is considered by many as the industry standard standalone software for detailed and accurate acoustic analyses, whereas Pachyderm is a plugin for Grasshopper (Rhinoceros 3D), which provides architects and engineers with more rapid feedback on preliminary design ideas because of its integration with a parametric modelling environment. Although an experienced acoustic designer could rapidly learn to use any of these programs, their usefulness during conceptual design or teaching activities is limited by the computational power and time required to get acoustic performance results. In these instances, testing time is more valuable than accuracy. This paper introduces Aeolus, an acoustic modelling plugin for Grasshopper. Unlike other similar Grasshopper plugins, Aeolus allows users to control the simulation accuracy and can therefore be used to rapidly test the performance of design ideas. Aeolus can also easily be interfaced with various Grasshopper optimisation plugins. Aeolus v0.1 was publicly released on 7 March 2023, and will be the focus of a Masterclass offered at the IASS Symposium 2023 in Melbourne, Australia.

This paper presents the development and application of a computational design tool that can be used to explore an AI-generated design space for the conceptual design of shell and tensile structures. An AI model was trained to extract geometric features from a dataset of 40 well-known design precedents of shell and tensile structures and to construct a design space. The trained model was then endowed with an interface to allow the designer to explore the design space within CAD software. Unlike the majority of current approaches to parametric design and optimisation, the exploration of the design space – and therefore the interaction between the designer and the computational model – does not take place via design variables, but through visual input. The potential of this tool to support the conceptual design of shell and tensile structures is examined through an application involving iconic design precedents. The application shows that, unlike form-finding and optimisation, this tool generates design suggestions that are not performance-driven, and do not require the statement of the boundary conditions, which would pre-determine the results. Despite this, such design suggestions can be considered plausible because they embed specific design knowledge resulting from a re-elaborating process of the main geometric features of the precedents used to train the AI model. These features include, for example, the shape of the openings, the number and location of the support points or the inversion of curvature, where present. The application results question the role of computational tools in conceptual design and illustrate an alternative strategy to explore the design space.

Biodiversity is in a state of global collapse. Among the main drivers of this crisis is habitat degradation that destroys living spaces for animals, birds, and other species. Design and provision of human-made replacements for natural habitat structures can alleviate this situation. Can emerging knowledge in ecology, design, and artificial intelligence (AI) help? Current strategies to resolve this issue include designing objects that reproduce known features of natural forms. For instance, conservation practitioners seek to mimic the function of rapidly disappearing large old trees by augmenting utility poles with perch structures. Other approaches to restoring degraded ecosystems employ computational tools to capture information about natural forms and use such data to monitor remediation activities. At present, human-made replacements of habitat structures cannot reproduce significant features of complex natural forms while supporting efficient construction at large scales. We propose an AI agent that can synthesise simplified but ecologically meaningful representations of 3D forms that we define as visual abstractions. Previous research used AI to synthesise visual abstractions of 2D images. However, current applications of such techniques neither extend to 3D data nor engage with biological conservation or ecocentric design. This article investigates the potential of AI to support the design of artificial habitat structures and expand the scope of computation in this domain from data analysis to design synthesis. Our case study considers possible replacements of natural trees. The application implements a novel AI agent that designs by placing three-dimensional cubes – or voxels – in the digital space. The AI agent autonomously assesses the quality of the resulting visual abstractions by comparing them with three-dimensional representations of natural trees. We evaluate the forms produced by the AI agent by measuring relative complexity and features that are meaningful for arboreal wildlife. In conclusion, our study demonstrates that AI can generate design suggestions that are aligned with the preferences of arboreal wildlife and can support the development of artificial habitat structures. The bio-informed approach presented in this article can be useful in many situations where incomplete knowledge about complex natural forms can constrain the design and performance of human-made artefacts.

This paper presents a theoretical framework for the implementation of Artificial Intelligence (AI) in architectural and structural design processes, and it is complemented by some practical applications. The aim is to demonstrate that AI can be used to simulate certain aspects of human cognition and can therefore be integrated into CAD software to support conceptual design and idea generation in a number of different ways. The aim of this study is also to investigate to what extent AI models can interact with a designer to explore future forms of human–machine interaction, including autonomous and participative design. This study identifies and applies AI models to simulate three distinct learning mechanisms: design expertise, playfulness and analogical reasoning. Each strategy has been applied to train different AI models, including generative models and reinforcement learning agents. In the first application, the AI model extracts visual features from a dataset of shell and spatial structures, and then recombines such features to generate new design propositions. In the second application, an AI agent learns a design strategy to solve a toy-design problem with no prior knowledge of precedents. The third application illustrates that AI can be trained to discover meaningful features from biological forms and generate simple design objects through the visual abstraction of such forms. The applications demonstrate the ability of AI to synthesise design options and interact with a designer through visual data formats, such as 2D images and 3D models. This work does not focus on assessing the usefulness of AI models in a real-world design scenario, or on comparing AI with current computational design tools and approaches. It instead investigates different forms of design exploration for computational design purposes, thus paving the way for the development of future autonomous and participative design systems.

This paper presents a comparison between human-defined and AI-generated design spaces through simple optimisation applications.

A design space is a formal expression of a design idea. It is constructed by selecting a set of variables, which limit the search for suitable solutions to a design problem within a specific range of options. Most computational approaches to structural design are based on parametric modelling, which require the definition of a design space, and therefore an analytical formulation of a design idea. In structural optimisation, such approaches tend to limit the search for optimal solutions to a subset of the entire space of design possibilities, and do not necessarily prompt the designer’s creativity.

Recent AI models, such as Variational Autoencoders (VAEs) (Kingma and Welling, 2014), have the potential to overcome some of the limitations described above. VAEs can construct design spaces by extracting implicit design variables from a dataset of design solutions. Such variables result from a learning process and are conditioned exclusively by the characteristics of the dataset, rather than by a human-formalisation of design thoughts.

A VAE has been trained on an artificial dataset of shell structures to construct a design space, which has then been compared with a design space constructed through the explicit definition of design variables. The comparison has been performed by analysing the diversity of the solutions retrieved from both design spaces in two optimisation applications.

The comparison demonstrates that optimisation based on AI-generated design spaces results in a greater diversity of design outputs than the predictable solutions provided by optimisation based on human-defined design spaces. Furthermore, such design outputs respond better to the selected performance criteria.

‘Sketches of Thought’ is a human-machine collaborative design system based on Artificial Intelligence (AI).

The first aim of ‘Sketches of Thought’ is illustrating an approach to AI-integration within the designer’s creative workflow. The system translates a hand-drawn architecture sketch into a photorealistic image that suggests a possible evolution of the design idea. The dialogue with the system happens through a visual interface whereby the designer communicates by sketching directly on a drawing monitor, while the system responds by showing the results of the image translation process on a second monitor. Interaction with the system does not end after a first iteration. Instead, the designer is encouraged to adjust the initial sketch – or even make new sketches – for several times to explore, with the aid of the machine’s feedback, different elaborations of an idea. This system does not require particular drawing skills, and therefore anyone can experience a proficient ‘exchange of ideas’ with the AI model.

The second aim of ‘Sketches of Thought’ is helping the designer familiarise with AI technology. This is achieved by unveiling the black box of the AI model functioning, that is, through a representation of its internal processes. Moreover, as the AI model simulates some aspects of human cognition, a look inside the black box of AI also means visualising a simplified version of the human mental processes. Therefore, learning about AI is an opportunity for humans to learn more about themselves.

The relevance of this virtual prototype is twofold. First, it promotes the view of AI as a means to augment rather than replace the human cognitive capabilities. Second, it challenges the current beliefs and prejudices on AI-technology by making the AI internal processes explicit through a visual representation.