During his 2015 TED talk “What happens when our computers get smarter than we are?”, Swedish philosopher Nick Bostrom claimed that machine intelligence will be the last invention that humanity will ever need to make [1]. Artificial Intelligence (AI) has vastly revolutionised industries such as finance, healthcare, marketing, and information technology [2]. But how has it affected the architecture and construction industry? Will it play a role in aiding architects to spend more time designing or will it eliminate the designer as middleman?
Upon the release of text-to-image software such as DALL-E in 2021 and Midjourney in 2022, the potential for AI to impact the creative industry has become a reality. This has brought both fear and curiosity to the forefront for architects. So far, text-to-image software has caused significant controversy within the field of art due to a potential loss of client base for artists. With a text prompt, any user can generate artwork within seconds, though many have debated whether the work could ever be considered “art”. Additionally, the use of an artist’s name within a text prompt to generate images in their style has caused subsequent issues of copyright infringement [3]. On the other hand, it has been argued that the general use of text-to-image software to augment existing practices has enabled artists to speed up the process of producing concept work and enabled designers to communicate with their clients in understanding their ambitions [4]. The creation of images in this way for architectural purposes could be beneficial for the architect during the early stages of design. However, use of text-to-image software would not be considered useful by many past these stages due to the inability to translate these AI-generated images into detailed construction drawings. As with concept sketches in a notebook, the architect must still use their knowledge to produce the final project outcome [5].
However, it is still possible that clients may turn to AI for reasons of cost and efficiency in the commissioning of new buildings. To use an example that might concern architects: it is already common practice in Ireland for the Department of Education to procure school buildings using a generic repeat design to quickly produce buildings [6]. This can be partly attributed to the 2008 recession and baby boom, which prompted the need for schools both quickly and with a focus on cost. When we consider the use of AI and its ability to utilise deep learning to produce work at such a fast pace with limited cost concerns, it could potentially be an efficient way for government bodies to procure buildings without the requirement of an architect as an intermediator. Similar to how text-to-image software learns from existing art and photography to generate recomposed images, it is highly achievable for AI to generate a floor plan utilising precedent layout drawings of existing buildings. We must consider how this would impact the architect's role in the built environment if the art industry is any indication. How might we deal with copyright infringement should building designs be generated by AI using references to the work of existing architectural practices?
Contrary to the belief that AI will mitigate the role of the architect, many have argued that AI should not be the enemy, but rather the liberator, enabling the architect to be more human at work. Adel Zakout, of furniture-sourcing platform Clippings, has claimed that in the coming decade designers will benefit from AI by utilising it to perform admin tasks within the office, thus allowing more time to create [7]. In addition to this, the deep learning of AI could be extremely beneficial to the architect in reviewing designs under the scope of building regulations or other desired parameters. This potential could limit human error within the design process. We are already seeing real-time use of AI in this way. ‘Architectures’, an AI-powered building design web tool is already working on a process whereby the software has been trained to fully adapt to specific building typologies and design rule specifications [8]. It generates building typologies within pre-set parameters, customising a bill of quantities and financial planning and the integration of BIM. Created by Smartscapes Studio, they claim that this software intends to cooperate with the user, utilising AI to enable the user to speed up the development process of a project significantly [9].
We have also seen the use of AI within parametric architecture, aiding in the development of more complex forms, light analysis, and environmental efficiency. Practices such as Zaha Hadid Architects have already begun utilising AI in their projects to determine both form and optimise building performance [10].
The use of AI within the architecture and construction industry could be seen as a double-edged sword. With the potential for it to reduce the duration of admin tasks within daily practice, it is no wonder that some architects have begun to utilise this new technology in practice. However, whether it will be advisable for architects to rely on AI to regulate or design their project is debatable as it could lead to potential claims of negligence or a loss of knowledge within the profession. Be that as it may, the integration of advanced technologies into daily practice is inevitable. Its potential to generate complex forms, optimise lighting design, and environmental efficiency means that AI platforms could be regarded as game changers for the built environment. We should consider their potential proactively rather than fearing their use. Architects should not see them as replacements but rather aids that could enable them to be more human in the workplace.
Present Tense is supported by the Arts Council through the Architecture Project Award Round 2 2022.
1. N. Bostrom, ‘What happens when our computers get smarter than we are?’, Nick Bostrom, [online video], 2015, https://www.youtube.com/watch?v=MnT1xgZgkpk, (accessed 10 February 2023).
2. A. Jarra, ‘The Future of AI: 5 Industries That Will Be Most Affected’, GetSmarter, [web blog], March 09 2022, https://www.getsmarter.com/blog/market-trends/the-future-of-ai-industries-that-will-be-most-affected/, (accessed 10 February 2023).
3. K. Chayaka, ‘Is A.I. Art stealing from artists?’, The New Yorker, [website], February 10 2023, https://www.newyorker.com/culture/infinite-scroll/is-ai-art-stealing-from-artists, (accessed 10 February 2023).
4. B. Dreith, ‘How AI Software will change architecture and design’, Dezeen, [website], 16 November 2022, https://www.dezeen.com/2022/11/16/ai-design-architecture-product/, (accessed 15 February 2023).
5. B. Dreith, ‘How AI Software will change architecture and design’, Dezeen, [website], 16 November 2022, https://www.dezeen.com/2022/11/16/ai-design-architecture-product/, (accessed 15 February 2023).
6. T. Sheppard, 'Ireland's Generic Repeat Design Schools Programme', 1 September 2011, CELE Exchange, Centre for Effective Learning Environments, no. 2011/05, OECD Publishing, Paris.
7. M. Fairs, ‘Artificial Intelligence "will empower designers" says Clippings co-founders’, Dezeen, [website], 19 August 2021, https://www.dezeen.com/2021/08/19/artificial-intelligence-empower-designers-clippings-co-founders/, (accessed 15 February 2023).
8. ‘Product’, Architechtures, [website], 2020, https://architechtures.com/en/ (accessed 15 February 2023).
9. ‘Product’, Architechtures, [website], 2020, https://architechtures.com/en/ (accessed 15 February 2023).
10. J. Stuhlinger, ‘AI in the desert by Zaha Hadid Architects’, UBM Magazine, [website] https://www.ubm-development.com/magazin/en/intelligent-dunes-by-zaha-hadid-architects/, (accessed 15 February 2023).
Website by Good as Gold.