Imagine you live in an apartment building. But you did not choose the building manager, nor do you know the concierge who shares the hallways with you. All the condo rules—from when the lights turn on, how the elevators operate, the use of security cameras, and the distribution of water—are defined by a foreign company you have never met in person. This company collects data about your habits: what time you leave home, who you talk to in the elevator, and even whether you slept well—based on how you turn the doorknob.
One day, you discover that all this information, including intimate data, ended up in the hands of a pharmacy that bought your “profile.” Your schedules, conversations, routine, even pictures of you, were all leaked, sold, or used in decisions you never approved. A neighbor tells you they received a strange ad for medication based on information from their smart fridge. And for weeks, someone in the building has been saying the “algorithm” predicted you would not pay next month’s rent.
You are outraged—and rightly so. You would probably think about calling the police, taking legal action, requesting a tenant meeting, demanding your rights. But none of that is possible in this building.
The building manager lives in another country, because in fact, he is a supervisor managing dozens of buildings at once using an “algorithm.” And by contract, the company that runs everything is not subject to local laws. It cannot be summoned here, does not attend meetings, and only communicates via app chatbot notifications.
What might sound like an episode of Black Mirror is not far from what is already happening in our lives thanks to artificial intelligence (AI) systems operated by big tech companies. They already regulate what we see, what we consume, what we think we choose. And the most troubling part is that this regulation is not carried out by any public or democratic body—it is done privately, opaquely, and remotely.
AI is already regulated—by those who control it
Platforms like OpenAI, Meta, Google, and Microsoft are constantly making decisions about what their systems can and cannot do. They define, on their own, what is considered dangerous, unacceptable, true, or false. Based on obscure criteria, they decide what content gets visibility and what gets hidden, which uses are permitted and which must be blocked. These are high-impact decisions, made centrally, without any democratic process. They are not subject to public consultation, do not adhere to Brazilian legislation, and fail to consider the cultural, social, and economic contexts of our country.
These companies often claim they follow ethical guidelines and that the technology is used responsibly. However, the actual governance of these platforms is limited. They create ethics and responsibility advisory boards—but with fictional, often purely symbolic power. They publish impact reports—but these documents are written by the companies themselves, with little transparency about methodology and without independent audit mechanisms from civil society or local regulatory authorities.
While some still associate state regulation of artificial intelligence with ideas like censorship, obstruction, or limits to innovation, the truth is that AI is already being regulated. It is shaped by those who design it, fund it, and define its technical, commercial, and ethical parameters. In practice, it’s like living in a condo under laws created by foreigners—with no voice, no vote, and no recourse.
Legal regulation is not enough: Infrastructure investment Is essential
The contract signed between Microsoft and Brazil’s Federal Government from October 26, 2023, to January 31, 2025, totaled R$1,272,555,777.49 (about 230 million USD). This amount refers only to what was paid by the Executive Branch, not accounting for contracts with other entities, the Public Prosecutor’s Office, and other public institutions. Part of this investment was allocated to cloud computing services—critical infrastructure often used as the foundation for systems that mediate the delivery of public policy services. In other words: if we don’t pay the rent with the taxes we collect, public policies could be impaired, since services like Microsoft’s are what enable these public policies to be offered in the country.
To make matters worse, the Brazilian government—through its Finance Minister, Fernando Haddad—presented a proposal in early May 2025 to exempt big tech companies from taxes on the construction of data centers in Brazil. On top of paying an exorbitant rent to live in a building that is not ours, we also advocated for a generous property tax discount for the landlord. In practice, we remain without access to the keys, without decision-making power, and still benefiting those who profit from our dependency.
This contract, among many others with major tech companies, reveals Brazil’s deep technological dependence. We’re paying a high price for technologies we did not develop, do not control, and that often do not reflect our social, economic, or strategic interests. It’s an asymmetrical relationship where the country acts as a passive client, while technical and commercial decisions continue to be made abroad.
Given this scenario, regulation is urgent and non-negotiable. Bills like 2630/2020, which seeks to hold digital platforms accountable, and 2338/2023, which regulates the legal framework for artificial intelligence in Brazil, are key steps toward establishing public limits and guaranteeing minimum rights in the use of these technologies. However, it’s important to emphasize that this is just part of the path—not the endpoint. Regulating the use of AI—mostly foreign—in Brazil is like monitoring what others are doing in our own backyard.
However, what we want goes far beyond that. We want our own building, with architectural design based on our reality. We want a sustainable Brazilian data center, a Brazilian and plural AI, science and technology with ongoing funding, centered on our needs and values. Not just AI from the United States or China. What we need is technological sovereignty with infrastructure investment—so Brazil is not just a client or a digital colony, but the protagonist of its own transformation. Otherwise, what will the next rent adjustment be?
Latinoamérica21, together with The Conversation Brasil, Brasil de Fato, and other allied platforms, promotes—through collaboration with Brazil’s National Network to Combat Disinformation (RNCD), Ibict, and ICIE—the dissemination of content aimed at fostering a more informed and critical citizenry, to confront disinformation, a growing threat to democracy, science, and human rights.
Ergon Cugler holds undergraduate and postgraduate degrees from the University of São Paulo (USP) and a master’s degree in Public Administration and Government from the Getulio Vargas Foundation (FGV). He is an associate of the postgraduate program at the University of Barcelona and a CNPq researcher affiliated with the Laboratory for Studies on Information Disorder and Public Policy (DesinfoPop/FGV).
*Machine translation proofread by Janaína da Silva.
Autor
Graduate and postgraduate from the University of São Paulo (USP), master in public administration and governance from the Getulio Vargas Foundation (FGV). Associated with the post-graduation program of the University of Barcelona. CNPq researcher linked to the Laboratory of Studies on Informational Disorder and Public Policies (DesinfoPop/FGV).
