The Structure Determines the Order: Taking a Stance in Architecture
Recently, the wave of boycotts in the architecture and design community has reminded us of a bigger truth: Architecture has never been just about designing buildings.
Every decision that shapes cities and structures directly affects the social order, economic balances, and people's lives. Therefore, thinking that an architect can remain detached from politics is contrary to the very nature of the profession. A building is not just about walls, columns, or aesthetic concerns. It gains meaning through the land it stands on, the lives it touches, and the culture it shapes within. That is why the planning of public squares, urban renewal policies, social housing projects, or the rise of skyscrapers is not just about engineering or aesthetics. It is never just about negotiations between employers and contractors. These are reflections of a form of governance. Cities are mirrors of power; the prosperity, justice, and even ideology of a country can be read through its urban fabric. The increasing suffocation of poor neighborhoods, the transfer of public spaces to private capital, the concrete swallowing up of green spaces… None of these are coincidences. Each is a conscious choice and determines which segment of society will have what standard of living.
This is why architecture is political.
Throughout history, architecture has been one of the most visible manifestations of power. The monumental structures of the Roman Empire, the public külliyes of the Ottoman Empire, the rigid and orderly urban fabric of the Soviet Union... Each of these represented an ideology. For example, the idea after the French Revolution that cities should be opened up to the public was directly reflected in urban planning. In the 20th century, the Nazis’ monumental and authoritarian architectural approach was a tool for reinforcing the regime’s power. The Cold War rivalry between the United States and the Soviet Union even influenced architectural styles. Today, neoliberal urban policies shape our cities with massive shopping malls and luxury housing projects. This situation deepens social inequality, spatial segregation, and even social unrest.
The current state of the architecture community in Turkey is an extension of this. That is why boycotting a large design platform is not just about protesting a page or an institution. This boycott is an attempt to take a stance against the injustices, censorship, and monopolies within the system and sector that architects belong to. Since architecture cannot be separated from politics, architects cannot remain indifferent to social events. Every decision an architect makes while practicing their profession sometimes destroys a neighborhood, sometimes creates a living space.
Architects are the actors who shape a city’s memory, identity, and future. Every structure they build, every decision they make, affects who will lead what kind of life. The fact that a public building is accessible only to certain segments, the profit-driven execution of urban transformation projects, the reduction of green spaces... Each of these is a political choice. Therefore, claiming that architecture is apolitical is somewhat of an escape from reality. That is why today, our colleagues face a clear question: Will we view architecture merely as a technical endeavor, or will we act with an awareness of our social responsibility? The difference between being a passive observer and taking an active stance is not just a professional attitude but also a matter of conscience. Strength to your blocks!
Artificial Intelligence Simulating the Human Mind: Are We Ready for the Future?
A major breakthrough is occurring in the world of artificial intelligence. Recent research from Stanford University shows that AI-powered generative agents can simulate human behavior with astonishing accuracy. This study goes beyond traditional AI models, revealing a system capable of predicting and even precisely mimicking human emotions, thoughts, and decision-making mechanisms. While large language models are already proficient at generating human-like text, these agents internalize human behavior, modeling how individuals think. This technology has the potential to revolutionize social sciences and policymaking, yet it must be evaluated alongside its ethical and societal implications. But that’s not our focus for now.
Let’s go back to 2023. The first author of the study I mentioned earlier and their team published research that year, outlining an architecture for creating human-like interactive agents step by step. First, they developed a long-term memory system for each agent, capable of storing over 100 records. Then, the most relevant 100 memory records were retrieved to guide the agents' behavior. To make the agents act more coherently and realistically, a "reflection" mechanism was added, allowing them to generate higher-level thoughts from their observations. Each agent was programmed to engage in this reflection process approximately two to three times a day.
At the start of this simulation, everything began with an agent intending to organize a Valentine’s Day party. Despite multiple possible points of failure—agents might not act toward this goal, they could forget to notify others, or they might fail to remember to attend (how very human!)—the party eventually took place, and a group of agents gathered and interacted.
Here’s an illustration of a morning in the life of one such generative agent. The character John Lin wakes up around 6 AM, completes his morning routine—brushing his teeth, showering, and having breakfast. Before heading out for work, he briefly interacts with his wife, Mei, and their son, Eddy. Similarly, daily activities were structured into broad time slots (5–8 per day), with each slot divided into detailed sub-tasks lasting 5–15 minutes. During testing, a virtual town called Smallville was created, housing 25 different agents, whose social interactions were observed. A human evaluation study involving 100 participants assessed the impact of planning, memory recall, and reflection processes. The results showed that when all these components worked together, the agents exhibited the most realistic, human-like behaviors.
After this pioneering study, we can now turn to the new research I’m eager to discuss: Generative Agent Simulations of 1,000 People.
Joon Sung Park and his team conducted interviews with 1,052 people, asking them to talk about themselves for two hours—detailing their daily lives, beliefs, and values. The researchers processed these deep-dive interviews through AI to create generative agents that modeled individuals’ characteristics and behavioral patterns. These agents were then given different scenarios to predict how their real-life counterparts would respond. But the real challenge began after that: the agents were expected to act in ways that matched human behavior. In other words, they would develop their own personalities based on environmental conditions. (If you’ve ever reflected on human behavior, you know that it is rarely predictable, and strict rules cannot fully define it. That’s why only an AI-driven system can successfully replicate human behavior.) The results were striking: these agents predicted human responses with 85% accuracy. In other words, just as people maintain consistent views over time, these AI systems could make predictions with similar reliability.
Even at this stage, it’s incredible!
A similar study was conducted by Robert Yang and his team: Altera: The First Agent-Based Civilization. This research created an environment where 1,000 agents could use natural resources, develop economies, and even form their own cultures and religions. (You can watch a video on Yang's study here.) Initially, these Minecraft-like agents had nothing, but they established marketplaces using found resources, engaged in trade, and even created gemstones to use as currency! (Lydians, who?)
Now, if this study with 1,000 agents already provides meaningful insights, imagine what a version simulated with millions of agents could offer. This is the creation of a human-centered world! These artificial agents could generate countless new industries, particularly in culture and economy. In the future, we may see AI-driven marketing services or entire business sectors run by generative AI agents. These agents could create their own currency, develop economies, and sell their products at remarkable profits!
These advancements will revolutionize areas such as public policy, marketing strategies, and human psychology. Governments will be able to simulate public reactions to specific legislation before implementing it, while companies will no longer rely solely on surveys but instead use AI models capable of making highly accurate predictions. The fascinating part is that these systems won’t rely on demographic stereotypes; instead, they will analyze each individual’s unique attitudes and behaviors to make predictions! AI-generated human simulations may even serve as virtual laboratories for testing societal reactions to major events.
All of this brings forth a fundamental question about human intelligence and decision-making: If individual political preferences, shopping habits, or ethical choices can be predicted with high accuracy, does this challenge the uniqueness of human will? Are the decisions we believe we make consciously actually just the outcome of an algorithm? The level AI has reached today is not just an engineering marvel—it also raises profound philosophical and ethical questions.
In the coming years, these AI systems will continue to evolve and become an integral part of daily life. Companies will use them to better understand consumer needs, and governments will incorporate them into policymaking. However, questions about regulation and individual privacy must be addressed in future discussions. If we view this technology purely as an engineering feat without considering its ethical implications, it won’t be difficult to foresee the risks it may pose in the future.
AI is no longer just a tool for processing information; it is becoming an entity that understands and simulates human behavior. At the end of this cycle, perhaps these AI agents we’ve created will begin generating their own simulations of us (an Ouroboros story). Who knows?