In 2022, a wider audience gained access to unexpectedly powerful AI tools, including Stable Diffusion, Midjourney, and DALL-E 2 for text-to-image generation, as well as the human-like chatbot OpenGPT.
One year later, pundits, organizations, and governments have stated these technologies will pose profound risks to society and humanity—from automation-spurred job loss to disrupting democratic processes to the automatization of weapons.
In July, Google, Microsoft, and OpenAI founded the Frontier Model Forum to regulate AI development. Meanwhile, in November, 28 governments signed the Bletchley Declaration, “the first international declaration to deal with the fast-emerging technology” acknowledging AI potential "catastrophic risk to humanity.”
As part of the latest ArchDaily’s Year in Review edition, we asked ChatGPT for predictions on 2023's architecture trends. One key takeaway was that OpenAI’s chatbot “can provide information and examples based on the descriptions it has read, rather than providing its aesthetic analysis”—at least for the time being. With this interview, we wanted to encourage architects and designers to embrace these technologies for our good in the early stages of development, instead of later suffering the effects of disruption of employment models implemented by someone else.
Prompters and architecture bias
When Alejandro Aravena’s ELEMENTAL released the blueprints of four social housing projects for open-source use in 2016, the idea was that architects and the general public worldwide could learn from them. However, beyond architects, only a few possess the technical expertise to read the plans, while even fewer can afford to build them. Similarly, ChatGPT and, by extension, any AI technology, would not replace humans but will overly empower those who know how to prompt—and those who own it, naturally.
On one hand, recent disruptive technologies led to the mass creation of AI-generated imagery, sidelining design professionals (illustrators, designers, and photo stock companies). On the other hand, they give rise to a new specialization of undetermined duration: prompters—individuals skilled in prompting. Text-to-image generations are a great example: you can look at great results on Reddit leading to think that anyone can easily achieve that level as well, but creators hardly will share their prompts—that’s creators’ value.
Midjourney and ChatGPT’s knowledge has been acquired by reading the data of millions of websites, thus, both the generative program and the chatbot’s training reflect the current status of the internet data. Since 63.7% of websites on the internet are in English, according to Statista, it makes sense that ChatGPT “is skewed towards Western views and performs best in English" as OpenAI recently acknowledged.
The architectural knowledge of these technologies mirrors the same bias: Midjourney images based on prompts like "[typology] designed by [architect]" from first-world architects produce more aesthetically accurate results. This discrepancy is not due to colonialist or racist biases but rather because the model's crawled data regarding these architects is bigger allowing better results than when asked to imitate specific Mexican, South African, or Indian architects’ styles.
Predictable architecture
Artificial intelligence is trained by humans but exceeds our capacities as it processes vast amounts of data, identifies complex patterns, and makes decisions based on statistical probabilities. As built architectural production heavily relies on predictable processes that can optimize budget, physical and human resources, speed up construction timeline, and achieve faster scalability of companies’ commercial operations, then the chances of automatizing most architectural production are high, especially real estate projects.
AI can even optimize new processes by identifying new patterns we do not know yet. However, not every decision on an architectural project’s timeline is predictable or efficient. Aesthetics, market trends, marketing campaigns, general public opinion, and stakeholders’ interests —namely, clients, developers, architects, and managers— have always been part of the equation. As long as humans are the ones who make the final decision, then AI will be subordinated to ordinary decisions.
Then, the question has to be asked: Will AI technologies replace good architecture?
While Singapore and Dublin have launched their digital replicas using machine learning to predict future events and trends, entire countries have not ensured clean water and electricity for their inhabitants. Since architecture reflects societies (and not the other way around), the accentuated social inequality we live in will continue to be reflected in the architecture we build: some works entirely designed by artificial intelligence, others crafted by hand and physical models in an architect's boutique office, and the vast majority made on-site with paper and pencil without architects on the middle. Perhaps all three scenarios exist within the same city.
Writer Benjamin Labatut stated:
If artificial intelligence were to think, it would have blind spots; if it manages to be creative, it will have limits, because limits are fruitful; if it is capable of imitating our capacity for reasoning, it might need (or develop) our talent for madness. And if it lacks understanding, if it doesn't care about the beauty and horror it can create, then it would be foolish to put ourselves in its hands.
The future of architecture lies at the intersection of technological innovation and human intent. Ultimately, human agency—civil society, politicians, and stakeholders—holds significant influence. The course of history is not written in stone, but shaped by decisions made today, especially if AI affects our pockets.
Architecture, then, becomes a result of collective decisions, where the advancements of AI intersect with the aspirations and values of society. It's within this interplay that the evolution and impact of architecture find their resonance and significance.
This article is part of the ArchDaily Topics: Year in Review, presented by Randers Tegl.
"When creating unique architecture, visionary ideas aren’t always enough. A unique look demands character, courage, and distinctive materials. And a format to achieve the extraordinary. At Randers Tegl, we aim to add a unique touch to exceptional brickworks by bringing premium bricks to life and into the world of architecture. Making the impossible possible. We are proud to be a part of unique architecture worldwide since 1911."
Every month we explore a topic in-depth through articles, interviews, news, and architecture projects. We invite you to learn more about our ArchDaily Topics. And, as always, at ArchDaily we welcome the contributions of our readers; if you want to submit an article or project, contact us.
