The City That Watches Your Daily Life

You may have also noticed that new offices have thermostats that know exactly when you’re coming in even before you do. They know that you arrive at 8:57 a.m. on Monday mornings, that you grab a coffee around 9:00 a.m., and that you prefer a comfortable temperature by 9:30 a.m. The system dims the hallway lights and preheats the conference room you haven’t booked yet. This isn’t a futuristic novel; it comes from a product brochure I recently read about smart office management.

The problem is that this fascinates us as we read about it, because we no longer have to press so many buttons or plan so far in advance before we get home or to the office. But do we really notice what’s happening to our cities as we begin to automate everything?

AI as The New City Administrator

We’ve been told that everything related to AI in our lives today is called “progress.” So the building breathes, and the city adapts. And this infrastructure becomes an organism. Many people today speak of the “urban brain,” an AI layer hovering over every city, with thousands of sensors and data streams that control traffic flow, energy prices, weather forecasts, and license plate recognition. And this AI coordinates all of this in a seamless, optimized workflow.

For many cities, this sounds too good to be true on paper, because it allows for better control of energy consumption, reduces traffic congestion, and enables the prediction of elevator failures even before the cable snaps. Who could possibly object to efficiency, or who could possibly object to a building that heals itself?

Let’s consider this answer in the context of a society shaped by agent-based AI, where AI agents become our colleagues or autonomously handle all the processes necessary for our daily lives. Opting for an AI-powered system for city administration means not only that you no longer have to wait for someone to come into the office or call the company, because the AI can now handle that for you. It means less downtime, longer equipment lifespans, and lower costs. When you extrapolate this to the entire city with its bridges, water pipes, and substations, the impact is enormous. A city’s entire energy supply can be optimized using an AI system that knows the weather forecast and the price of the regional power grid, precisely calculating how much energy the buildings need. This type of coordination can lead to measurable savings in both costs and CO₂ emissions. This makes it easier to achieve the long-awaited environmental goals for mitigating climate change. And for cities with heavy traffic, AI-based systems take traffic management to the next level by using an AI system that analyzes intersections in real time, adjusts traffic lights and their timing, and reroutes buses.

Each of these applications makes sense on its own. Each solves a real problem with a plausible technical solution. So the question is by no means whether a particular optimization works. The question is rather: What will become of the system once it reaches a state that is practically unmanageable for a city with, say, 9 million residents?

A Future in The Surveillance City

The same sensor array that reads vibrations in the pipe can also read the rhythm of your life. So, the occupancy sensor knows when you are at your desk and when you leave. The cameras that count pedestrians for traffic planning also track which faces of which human beings appear, how often, and in what company they enter in the morning. License plate readers can build a detailed map of every car movement across the city you live in.

The technical architecture does not distinguish between optimization and surveillance. It involves the same data, but it is processed with different intentions. And this is not a hypothetical problem: cities in the United States and Europe have already terminated contracts with providers of AI surveillance systems after civil rights groups exerted pressure. These groups had documented that systems originally intended for traffic safety had been quietly expanded to include law enforcement, immigration control, and the monitoring of protests.

So we see a system and a pattern that follows a consistent sequence. First, it gains a foothold under the guise of public safety or energy efficiency, and once the sensors are installed and the data is flowing, interest in secondary uses begins to grow. And then the most important question arises: Who owns this “brain,” including the city’s data formats, model architecture, API layers, urban traffic scenes, energy grid, and building management systems? All of this is typically bundled by a single provider or a consortium of technology companies. The city pays for the optimization, but the provider retains the intelligence. In many cases, it is assumed that only one provider owns the trained models, historical patterns, and predictive capabilities that make the system valuable. This is often referred to as the “Enclosure of Urban Intelligence,” a digital version of the land enclosures through which what was once a common good passed into private hands. The city becomes operationally dependent on a single platform that it does not fully understand and cannot abandon without losing years of accumulated institutional knowledge.

Let’s look for ways to address these risks, rather than rejecting a new technology that promises to transform our lives with a multitude of possibilities. Smarter buildings and better-managed cities are not a threat. The threat lies in the conditions under which this artificial intelligence is deployed in a city. Before implementation, all stakeholders in this process must define the following as matters of responsibility, not of technology: Who owns the data, who manages the model, who can terminate the contract, and where will the data be stored after termination? What we need is the work of researchers, urban planners, and civil society to develop agreements aimed at a kind of “architecture of accountability,” with principles that are neither radical nor incomprehensible. This contract and these rules are simply the application of an old democratic instinct to a new technological sector.

Open data formats and modular system designs enable cities to switch providers without losing the operational knowledge associated with the applications they use. Citizen-friendly contracts that define data sovereignty, data portability, and permitted use for the entire lifecycle of a building or infrastructure facility—and not just for the initial deployment—provide public institutions with legal recourse against creeping data extraction.

The Responsible Use Of a Collective Artificial Intelligence

Trials are currently underway involving civilian oversight bodies that review algorithmic decisions, mandate transparency regarding the training and updating procedures of AI models, and have the authority to suspend or shut down systems that are proven to cause harm. Overall, these are not yet fully developed solutions. This marks the beginning of a new era in which collective artificial intelligence will be managed responsibly over the long term. The key factor is the underlying paradigm shift, which shifts the question from “How can we make the city smarter?” to “How can we ensure that the city’s intelligence remains accountable to the people who live in it?”

The traffic camera will continue to count vehicles, and the heating system will continue to heat the rooms efficiently and gently in the morning, before you arrive at the office. None of this needs to be shut down, but the entire infrastructure that controls this artificial intelligence, with a framework and data architecture agreed upon with citizens, must be designed with the same care we once devoted to our old network of pipes and aqueducts for water supply.

We’ve done this before, with water supply, road construction, and electricity. The question is whether we have the time and patience to do it again, with a technology that is far less visible and is evolving at a breakneck pace, unlike anything humanity has ever developed.

If the city already knows your habits better than your neighbors do, at what point does it stop being your home and start being your landlord?