This article is the tenth in a series focusing on the Architecture of the Metaverse. ArchDaily has collaborated with John Marx, AIA, the founding design principal and Chief Artistic Officer of Form4 Architecture, to bring you monthly articles that seek to define the Metaverse, convey the potential of this new realm as well as understand its constraints. In this feature, architect John Marx questions the limits and capabilities of AI in architecture and in creating buildings that resonate deeply with people and communities.
The profession of architecture is often referred to as one of the world's two oldest. It has traditionally been the space of dreamers imagining new worlds and new ways of living. For much of this time, we have used machines to construct buildings and cities. As science has progressed from the mechanical to the digital enabling us to dream at levels beyond the contractor's toolbox, with new forms of construction, and with that came shifts in the design process. First as digital design with the use of 3D modeling, and now with the advent of AI the limits have expanded again.
While AI itself is relatively new, it is part of a centuries-old dynamic between humans and machines. We as humans tend to react with suspicion and /or elation relative to the introduction of new technologies. Fortunately, time has proven that most technologies provide substantial benefits to humankind. The success of this happens not with the introduction of the change itself, but rather in how we adapt to that technology and how we change the technology to better suit our needs. AI will be no different. As with most any paradox dynamic, it is a balance of both/and that provides the healthiest of outcomes.
In that spirit, let's explore the aspects of design that AI currently cannot master. This involves asking ourselves a provocative question: "What does it mean to be human?" We will find one part of the answer in the fundamental human element of care, which is something that architects at their best integrate as part of what they describe as their design intent. A capacity for empathy is one of the most important traits that differentiates us as humans.
Human Care and Emotion
In 1635 Descartes uttered those famous words "I think, therefore I am". For all the wonderous advancements we enjoy in science, law, and technology, there was an arrogance implied in that statement, ergo, that thinking is the highest order of human achievement. His statement misses a critical aspect of what it means to be human. It might read better like this: " I think; therefore, I am ..... I care; therefore, WE are." While AI can sometimes think at levels beyond human capacity, it currently is not capable of care.
Care might be seen as being emotionally engaged with an issue/condition, with a deep concern and involvement with the outcome. Here the outcome matters to you, "you have skin in the game", where a bad outcome could directly cause you pain and loss. Care is a critically important aspect of creating buildings that resonate with the public. But this sense of care needs to go beyond the perception of buildings into how they perform in society and environmentally. Cities and buildings that relate an emotion of care will in turn produce a stronger sense of "we", and stronger deeper bonds of community.
There is an opportunity here to reframe the way we are looking at the introduction of AI into the world. It would seem that we, in this current moment, look at this as a binary dynamic pitting Artificial Intelligence (AI) against Human Intelligence (HI), but again, as with Descartes, this is not the complete picture. Are we falling victim to seeing this as a binary paradox, rather than a both / and opportunity? I think going forward it might be illuminating to see what we might accomplish by embracing the concept of Artificial Intelligence with Human Intelligence and Emotion (E), or AI with HI+E.
Human Emotion is what enables us to feel the qualities of a space and to fall in love with places that resonate with us. First-hand experience is a critical aspect of an emotion-based design process, it is not the same as digitizing it, for example, knowing first-hand how a room might feel with a set of unique features because of the climate, the context, the materials used or its function are all fundamental to a well-rounded design concept. There is thus a significant difference between an architecture that is borne from the experience of building versus one that takes shape on a screen.
Judgement + Curation
AI is good at approximating pre-existing responses to language prompts in a way that is formal or pictorial, but it falls short in implementing the informal moves that come through the design process and how, for example, unexpected site constraints, surprising characteristics of a material, or emotional and cultural imperatives that can add a narrative layer to a project that profoundly bring a building to life.
In essence, AI lacks those visceral qualities and is only as good as its operator who directs a result based on a pre-determined catalog of information. While AI can be a strong partner in improving the design process and its outcome, humans are uniquely capable of judging and guiding the process to create emotionally resonant buildings.
The ability of a human designer to judge quality is paramount in a traditional design process. AI cannot judge the quality of an outcome. It cannot tell you where it sits in the history of architecture, or if the proportions are harmonious (only if they are following a set of rules), it cannot tell you if it is good, or most importantly if it is lovable.
AI can determine the likely popularity of a building based on the popularity of similar buildings, but not a specific new design effort. It is also critical to consider that popularity is only one design factor and is not the sole determinant of the best design course.
Another essential aspect of the traditional human design process is curation and design direction. Design direction involves picking the best ideas and formal expressions from a set of options. It also can be referred to as having "good taste". In recent years curation, which is a longstanding profession in the fine arts, has been applied to architectural design. A high-level definition from the 1800s still applies today: “The expertise and education to speak with authority.” Currently, AI cannot curate or provide design direction.
What AI does very well is to provide great value as:
- a collaboration tool
- as a generator of precedent studies
- as a generator of numerous variations on an idea
- as an idea sparker
- as an illustrator of concepts and form
It is an exquisite new pencil, to be used and adapted as a welcome member of a multifaceted team. In a classic design process, we craft multiple forms. Each formal idea contributes to an ultimate solution, within a messy process of navigating a balance of factors and formal opportunities. Who produces that design, a human or an AI program, begs the question: "Does it matter?". In the end, it is how we curate those ideas into a holistic building that is the important thing, where the ideas came from is of lesser importance to the crafting of the best formal response to a set of design conditions.
Vision
Vision is the ability to see and imagine how unknown design ideas can enhance the human condition beyond the obvious and predictable. At our best, architects have vision, we take design risks, we know when to push and when to yield, we can prioritize options making the critical value judgments that are so finessed and specific to a set of circumstances that AI is truly challenged in this context. Discernment in this way is a uniquely human trait. It is what leads to our most memorable buildings that achieve normal things in unusual and unexpected ways expressing a unique vision that makes an original contribution to architecture. It is this very originality that stems from a desire for humans to push forward that further distinguishes the power of HI+E (Human Intelligence + Emotion) from that of AI. Perhaps one of the most moving aspects of vision comes about when we reflect on how a human designer crafts a legacy over time, accumulating a collection of preferences, some deeply personal, others highly professional, yet often wholly interwoven in how they are expressed through the architect’s design intent.
Embracing New Tools and the Value of Being Human
This divide between architecture that is borne from a building versus one that takes shape on a screen is at the heart of conversations about what makes a lovable place. One that grounds people and communities, inviting them to take ownership of it. In their best moments, the human designer is full of quirky eccentricities falling towards the poetic suggesting that most things in life are part of an interconnected series of paradoxes that escape finely tuned algorithms.
If our collective goal is to create more human and resonant buildings and cities, we will need to deeply explore what it means to be human, how we can use our judgment and vision to craft environments people can fall in love with. In the last 60 years, we have moved away from the design of lovable buildings. We need to reconnect to what makes us human. AI can be a spark that helps us return to a more humane and caring world.
"AI and the Human Vector" is written by architect John Marx, AIA, the founding design principal and Chief Artistic Officer of Form4 Architecture, an award-winning San Francisco-based firm that designs prominent buildings, campuses, and interiors for Bay Area tech companies such as Google and Facebook, laboratories for life-science clients, and workplaces for numerous other companies. In 2000-2007, Marx taught a course on the topic of placemaking in cyberspace at the University of California, Berkley, and in 2020 he designed his first project in the Metaverse for Burning Man: The Museum of No Spectators. The following year, John Marx led a design team charged with creating a $500B portal to the Metaverse.
