AI Images are Ideological

Mother Nature is a pervasive theme in environmental conservation that assigns a human appearance to nature. The intention of this personification is to project the aspects of life-giving, female fragility onto nature in order to guarantee the survival of the human species. While God is the timeless alter ego for our ideological sense of ourselves, Mother Nature is the timeless alter ego for our ideological sense of nature. It is ideological because it is the fiction that is symbolic and specific to ourselves that we believe is necessary in order to survive with nature. It remains as an unquestioned ethical truth and yet, what results is not coexistence but a keeping up of appearances that re-represent nature as a preservable image. In the context of artistic production, keeping up appearances has meant continually chasing the creation of new, denatured objects with their own autonomy and rules. Just as the environmental conservationists want to escape death by telling themselves a fiction about nature, art culture wants to escape time into a higher state of consciousness by preserving novel art objects in museums and the institutions of taste. Like the big Other, Mother Nature, that does not really exist, the version of nature that artists want to produce is virtual. They want to create the virtual, the big Other, because that is what watches over and influences them.

Although making art does not have to be difficult, it usually is because artists never really know if they’re working too hard or not hard enough. The question of labor is central to producing novelty because artists must discipline their bodies and minds to be expert at their crafts which traditionally have taken a significant amount of time and devotion to achieve. While working and searching for new territories in their mediums, artists have been paranoid that their most authentic work may not be novel after all or, even worse, that they are merely copyists. But why is novelty so important in the first place? It may be that our species only chooses to preserve those things that are novel or have scarcity. I was once told during a review at SCI-Arc that I should be interested in developing scarcity in my work. I was perplexed by why I would ever do that. I thought, I shouldn’t need to artificially inseminate my work with rarity and besides, rareness does not equal novelty, the thing I am really interested in. A work of art can be completely banal and amateurish while still existing as a single, unique original, just as almost every mediocre, generative art NFT does on the crypto-marketplace. The advent of artificially intelligent methods of producing art have obviously reduced the labor pains of producing art objects but it is questionable to me whether they have produced novelty yet. After making that kind of statement, many of my contemporaries would probably shame people like me as luddites, unwilling to accept new technology but that shallow political critique misses the mark entirely. It is not the technology that I fear, it is the ideology of AI images.

Computing and ideology work well together because they are complementary. Artificial intelligence basically optimizes a fitness landscape of recorded data that measures difference and similarity between outputs of an ideological function. Sustainability technology does this well, where the goal is clear and the outcomes are measurable. Endless database fatigue is also becoming a thing of the past as AI collects and sorts information for us while our attention is shifted towards tailoring the managerial systems that structure information. Although, when the logic of AI optimization is applied to finding the fitness landscape of aesthetic novelty, we suddenly encounter a problem. To be human is to be flawed and to be flawed is to exist but it is unclear what a flaw means to artificial intelligence. As humans, we are supposed to feel guilt when we stray from ideology. Religion calls it sin and attaches the concept of forgiveness which never really works because it can’t undo time. Maintaining appearances in religion and society requires that people “forgive and forget” in order to keep on living but why do we forgive and forget the mistakes of fellow humans while not doing the same for AI? Perhaps it is because AI doesn’t feel guilt.

SCI-Arc initially approached AI images with the hope that there was some entropic element in the medium that could be revealed by disciplining the machine into producing productive failures. Could artificial intelligence generate an accident or is everything that artificial intelligence generates already an accident? These days, though, the internet has rapidly overtaken the pace of the school’s ability to dispense novelty by democratizing technology into computational models that anyone, including non-architects, can use. This occurred because AI image production now primarily uses methods of linguistic combination and transformation which inherently makes it a machine to copy what already is defined and exists. As AI’s accuracy and resolution got better, it became easier to judge whether a style transfer was good if it accurately depicted a subject with the style of another in high resolution. Within the formatting of images as data, sufficiently robust AI could reduce style to patterns, making it accessible to a wider audience, democratizing method while simultaneously devaluing its meaning. The pedagogical ramification of embracing style transfer as a labor to be automated signified a disinterest in the human construction of meaning through thinking and a preference for an accelerationist stance on rapidly producing combinations of style through codified language. Mesmerized by computational intricacy, we were like drug addicts trying to get a bigger, better high; one after another. Novelty was the drug of choice, injected through style transfer onto any cybernetic substrate we could get our hands on.

AI images are ideological because we are unaware that they are ideological. They are slowly redefining everything we once knew about art but like democracy, AI images have to be easy in order for them to be an effective ideology. What makes democracy ineffective is friction, tension, and difficulty which is never evident in the product but only in the process or method of creating art. I think the question that we keep forgetting to ask is if the meaning is in the how, not the what. That is to say, that method is the meaning. So when we displace the artist’s method onto automated machines, what is really left to work on? Perhaps, human agency is not what makes an artist creative and decision-making between predefined options is not agency, it’s labor. What or who made the rules in the first place? What or who made the style? Agency is constructing a new, unnatural object that disrupts the decision-making abilities of machines. Covid-19 had agency. The AI models trained on “normal” human shopping behavior simply broke when they encountered the sudden and extreme shift towards purchases of toilet paper, face masks, and hand sanitizer in 2020. The important difference, to me, is that the human artist intends to disrupt the flow of time by imprinting it with a moment when we had to pause and question if we were living in an ideology.

How to See Earth

I was recently watching a recorded presentation Gregory Crewdson gave at SCI-Arc a few years ago where he was talking about his artistic development and inspiration for photography after leaving graduate school. Like many recent grads, he struggled with feelings of immediate uncertainty so he moved to his parent’s cabin in Massachusetts to think and work in isolation. What he did there was never photographically developed but what he described was an intense period of working with the earth, constructing and photographing hundreds of dirt piles in his backyard. He called the piles totems, comparing himself to Richard Dreyfuss’ estranged character in Steven Spielberg’s “Close Encounters of the Third Kind,” who also experienced alienation from his family and community after crazily constructing his own piles, first from shaving cream, then mashed potatoes, and eventually dirt in his domestic living room. Artists, Crewdson said, pursue something outside their reach which makes that happen.

I initially came to know Crewdson through his staged photos of liminal suburban places in Massachusetts that were familiar to my Northeastern childhood experience. Those pictures work with the genre of domestic still life and he interestingly found ways to transplant domestic interiors onto suburban exteriors and vice versa, suggesting a moment after something dramatic had happened. I was always left pondering exactly what that was but I think the feelings of fear, distrust, and shame were very real and possibly there all along, even before Crewdson arrived with his film crew and set. He just showed us how to see them. While the eerily cinematic scenes Crewdson constructed may have been artificial, real acts of intransigence in the paranoid suburban communities of America do take place, though sometimes under the radar.

My parents, for instance, live in a suburban neighborhood of single-family homes in New Jersey and were recently fined over two hundred dollars by their local government for allowing wildflowers to spring up on their front lawn. I’ve dealt with similar people in planning board committees years ago when I worked as a practicing architect; they are the people who hold places of relative power while being the most distant from the impacts of their decisions. The distance between those public officials and the environment is critical to enforcing neoliberal order and subordinating the earth but it also enables certain people to exercise power over other people, even in the menial aspects of lawn care. Disciplining the lawn is really a way to represent white suburban enclaves while masquerading as responsibility towards communitarianism. You are not really punished for the wildflowers themselves but for not being one of us.

I think I pursued art and architecture because I always felt held back by my time and place, especially in the suburban neighborhoods of my youth, with their prim-and-proper lawns and crisply edged sidewalks. Wanting to be different, I took up street skateboarding early on. The competitive parks with vert-style half pipes behind fenced off regions in the local shopping mall never interested me because I felt that the existing suburban terrain was a more interesting place to dig into and discover and, besides, those were free public spaces I had a right to be in already. Like an artist who tried to expose opportunities in their medium, I would look for interstices where disciplining nature had failed; where the literal gaps, cracks, buckling sidewalk plates, and dilapidated pavement surfaces revealed a crumbly, organic matter considered unfit to tread on by any member of the community. Like standing off-center to view the distorted skull in Hans Holbein’s painting, “The Ambassadors,” skate-spotting required an anamorphic approach that could not be achieved by looking directly at the world. Now, as an adult and “professional”, I am confronted by a hegemony of code compliance officers, people who regulate the built environment vis-a-vis human accessibility and sustainability prescriptions, working in the most direct, neoliberal sense of judging a place to have positive or negative value. They discipline the streets and sidewalks, much like my parent’s onry, wildflower-police discipline front lawns, and subversively use the mechanisms of code compliance to subvert nature to a place of otherness.

One last anecdote to share is about a transfer student seminar I recently co-taught at SCI-Arc. The basic premise of the seminar was to instill an ethos of material upcycling in students, not for large scale infrastructures but at the scale of physical models that could be built in the short span of 3 weeks. They were challenged by material accountability, using only the volume of material cut out of the earth to fill reusable formwork while finding creative ways to combine earth, sand, stones, and recycled wood fiber board to grow totemic structures from their excavations. We found the existing precedents for rammed earth construction, which generally fell into the category of cheap construction for impoverished societies or the category of luxury residences that romanticize geological strata as a mute image, to be disturbing and useless to us. Instead, we took cues from Crewdson’s backyard dirt piles and the anamorphic approach of my skateboarding youth, searching for a different way to look at earth in the how, the method of doing and seeing, that would allow us to escape the limits of neoliberal aestheticization. By reorienting the position of our totem molds, we pre-cast earth into layers of deep color that evoked atmospheres of a multidimensional place. The textures were rough, sometimes chipped and broken, but that was just fine. We had no desire to cover up the earth because it was no longer just a mutable dirt, or matter out-of-place, but the entropic medium that all life on earth shares and cannot escape.

Can a Diagram Dance?

Typically speaking, diagrams are static representations. They don’t dance. These static representations are evidently subjective constructions and yet, pedagogy continues to legitimate them as objective, disciplinary tools. Even Charles Jencks’ “Theory of Evolution” diagram, that was an affectation of Deleuzian folds and surfaces which ironically mapped style and politics in a pseudo-scientific field of attractor basins, has somehow found a new life among activist-architects today who have interpreted its form as an objective diagram. The problem with diagrams, to put it simply, is not just that they are static but that they are dramatic. While it’s true that diagrams must work both within their formal and subject-matter’s rules, architectural diagrams have always wanted to do the impossible; to be both an object of beauty and a communicator of meaningful information. I take a stand that this is not possible and I disagree that we should continue to dramatize diagrams as a legitimizing device in schools. It would be wise if we could finally move beyond the incessant binary between artist and scientist that still irritates the discipline and profession but first, I should explain what is happening that led me to this conclusion.

Diagramming is a basic pedagogical tool taught in undergraduate and graduate schools of architecture that bases it knowledge production on the belief that there are object “types”' that exist as reduced parts of wholes, and the purpose of diagramming is to analyze those parts and their organization, or the elements and their forms, which are often represented as Euclidean point, line, plane, and volume geometries. Broadly speaking, generative design (also called parametric design) is an extension of this belief-system that is based on the gestalt principles of part-to-wholes that seeks balance and equilibrium. In a more traditional sense, the architect’s classic “napkin sketch” is a first attempt at representing a formal solution but it is also a diagram containing some additional information that its author intends to communicate. It can be read as a set of instructions, similar to a script, with the architect as the dramatist and the laborers as the actors who mechanically execute the diagram-script according to the architect-dramatist’s rules. This acting out of the diagram is a specific kind of ritual that might be called “practice” today. Architects, like doctors, practice their profession rather than just do it because there is a science underneath practice that continues to expand the breadth of disciplinary knowledge. Learning new skills and techniques can be an uncomfortable experience for beginners but both disciplines have encouraged their professionals to embrace discomfort for the sake of self-improvement or, worse yet, their “calling.”

Proficient professionals create efficient businesses but what efficiency really does is creates tension, much like drama creates tension, and never allows drama to dissolve back into dance. The attempt to make life better through dramatic architecture almost always leads to some form of tragedy. The tragedy of the corporate architect that lost their soul. The tragedy of the star architect that became famous and stopped thinking. The tragedy of the unpaid intern that wanted recognition but got exploited. Or, the tragedy of the architectural laborer that stopped believing in why they studied architecture in the first place. The saddest irony is that the majority of the world could care less about why a building looks the way it does and the universities could easily drop their diagram baggage, but the diagrams of metaphor and performance persist because they are so ingrained as an expression of the binary between artist and scientist that, I believe, will ultimately kill the discipline if it persists.

In schools, conceptual arguments for architectural form have been considered a legitimation technique for student evaluation that tests for diagrammatic legibility rather than complex, ambiguous thinking. Conceptual argumentation in architectural projects forces students to have “legitimate” reasons why they decide a building should represent some kind of metaphor, usually related to culture, or perform some kind of action, usually related to efficiency. Diagrams are easily consumed because the complexity of the world is reduced to simple lines that encapsulate an ideal representation. This traditional style of diagramming is a form of story-telling but what story is it telling? In spite of their expressive linework, gestural forms, and comical onomatopoeia, diagrams are still static objects that attempt to crystalize a metaphor about a single aspect of a building. This is problematic because it marginalizes the student and architect as peripheral stakeholders in a project, not invested in the dance of architecture as a participant in the ecology of planet Earth, but as a top-down world-maker that, in a best case scenario, fools their investors into believing that their metaphorical or performative diagram is somehow true. While the project on the metaphor has slightly waned in universities, the project on performance has not, as data has come to define the epistemic of evidence-based learning in the educational marketplace. The “sustainability” schools will argue that good diagrams explain a building’s environmental performance. Such diagrams represent building technologies, not really forms, that control solar radiation, attenuate noise, and collect rainwater. It’s important to recognize that these are all performances of drama, not dance. Considering that the environment is constantly changing, these diagrams are really just advertising to investors for building machines that subversively market sustainability by extending the limits of discriminatory human comfort amid climate change.

The drama of diagrams produces right and wrong answers because the metrics for evaluation can be critically narrowed to exclude all other possibilities and concerns, especially those outside of typical human and, more specifically, capitalist interests. If “Dancing about Architecture” means not being too dramatic, there really should be no good or bad architecture and, it should follow, that there are no good or bad diagrams. This brings me to my idea. What if there were a diagram that allowed ideas to coexist in a healthy ecosystem? A diagram that resisted affectations of form and pseudo-science. A diagram that discarded false part-to-whole relationships, metaphors, and efficiencies. A diagram that, maybe, isn’t a diagram at all. What I am proposing is a virtual debris field of dancing objects, a web environment that offers catharsis, a downward flowing energy that returns us to the earth, to an earlier state. I am proposing a diagram that dances.