Kant on a Tram
The Kantian readings and philosophical concepts in this text are not filtered through the lens of a philosophy professor, but rather emerge from curiosity started by questions that arose while watching the world from a tram window. Rather than a professional claim, it is just a thinking exercise about technology and architecture.
While watching outside the tram from its window, I tried to remember the lectures I listened about Kant. I hate to discuss AI but the issue goes much deeper than the mere capability of artificial intelligence to produce architectural forms. According to Kant, the conditions for human experience of the world include space as the form of outer sense and time as the form of inner sense: things are given to us in space, while our own inner states are given in time. Therefore, humans are not merely beings who think about space and time; our encounter with the world already takes place within space and time. Here, Kant’s (inverted) Copernican revolution becomes crucial: to understand how the world we experience appears to us, we must look not only at the objects themselves, but at the structures of the subject that make it possible for us to experience them.
Artificial intelligence can calculate spatial relations, process temporal sequences, analyze architectural representations, and generate new architectural images; yet the question of how space and time are “given” to it remains open. Its understanding of the concepts of “space” and “time” as I use them does not necessarily mean that these concepts constitute an experience for it in the same sense as they do for humans. Architectural representation must thus be rethought: whereas for humans, representation is the reconstruction of a world experienced in space and time in another form, artificial intelligence can produce new representations from representation itself. Therefore, the fundamental question is not “Can AI design architecture?” but a prior question: Must something first appear to a subject, given in space through an outer sense and in time through an inner sense, in order for it to be represented? If so, is the architectural image produced by AI truly a “representation,” or is it a reproduction of the forms of representation? Here, it may emerge that the difference between humans and artificial intelligence is not merely a matter of intelligence, computational power, or consciousness; more fundamentally, it is a matter of how the world is given to someone (or something?). Perhaps this is precisely the starting point for thinking about computational design and AI architecture: Humans first experience the world and then represent it, whereas artificial intelligence may be producing new representations of the world through existing representations. This distinction forces us to rethink architectural representation itself, the experience of space, and ultimately what we mean by “designing.”
Immanuel Kant (1724–1804) is one of the greatest thinkers of the German Enlightenment, who fundamentally altered the course of modern philosophy. Overcoming the long-standing conflict between rationalism and empiricism in his theory of knowledge, he established the system of “Transcendental Idealism,” arguing that the human mind does not passively perceive the world, but actively constitutes it through a priori (pre-experiential) forms such as space, time, and mental categories. In his moral philosophy, he laid the foundations of deontological ethics with the concept of the “Categorical Imperative,” which is independent of emotions or outcomes, resting instead on pure will and universal duty. With his three great critiques on knowledge, morality, and aesthetics (Critique of Pure Reason, Critique of Practical Reason, Critique of Judgment), he stands at the center of Western philosophical traditions.



