Artificial intelligence is changing the rules of the game — and we are not just talking about technology.
OpenAI just launched AI Futures, the official blog of its new team, Strategic Futures, created with a very clear mission: understanding how free society can reorganize itself in the face of an increasingly transformative AI. It is a small team, but with a massive goal summed up in a single thought-provoking question: how should free society be restructured to preserve individual rights and agency while accommodating the emergence of an AI capable of changing everything?
Within the community studying AI safety and policy, this type of concern usually gets a specific label: power concentration risk. And in the team’s view, this is the most serious, most challenging, and most difficult risk to solve in the long run. It is not alarmism. It is a concrete issue that is starting to show up in serious discussions inside the largest technology organizations on the planet — and the launch of AI Futures is yet another signal that this reflection has finally found an official home to grow.
And the question everyone should be asking is simple, but it carries enormous weight: who will have decision-making power in a world where machines can replace soldiers, bureaucrats, and workers?
Why power has always depended on people
Throughout history, political and military power has always relied on something very specific: human labor. Holding power means having a monopoly on legitimate force. And throughout the entire existence of humanity, that monopoly was made up of soldiers, what we now call police, and a massive civil and military bureaucracy. Sure, technological tools always helped in that process — but at the end of the day, everything was operated by human beings whose support was indispensable to keep the machine running.
And there is more: all of that apparatus — the soldiers, the tools, the bureaucracy — was funded by taxes collected from the wages and production of people who needed to be willing to pay. In other words, power, at its core, has always depended on large-scale cooperation and the consent of a whole lot of people. The philosopher David Hume summed up this idea well when he observed that political leaders have nothing to sustain them other than popular opinion.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
This implicit pact between those who govern and those who are governed was, for centuries, what guaranteed some level of balance — even if imperfect, even if full of flaws. Society always had a card up its sleeve: the ability to resist, to negotiate, to organize. But that balance may be reaching its end, and the emergence of highly capable autonomous systems puts this logic in question in a way we have never seen before.
What changes when AI takes over critical human functions
Advances in real-world autonomous systems could soon mean that states will be able to project force — both within and beyond their borders — without relying on soldiers or police officers willing to cooperate. And advances in artificial intelligence could allow the state to collect the revenue it needs not from the fruits of human labor, but from the outputs generated by data centers. Even bureaucracy itself could be largely automated.
In plain terms: power could stop requiring a deal with the broader society to sustain itself. And people, generally speaking, could end up with a tiny seat at the negotiating table — or maybe none at all. Imagine an army that does not need to convince soldiers to take the field, or a public administration that processes millions of decisions without a single human employee in the loop. This is not fiction. It is a technical possibility that is getting closer to reality every day, and it raises profound questions about what it means to live in a democracy when the tools of power no longer depend on collective consent.
Avoiding this scenario is essential to preserving human freedom. According to the OpenAI team, no scientific discovery, technological advance, or economic growth is worth sacrificing the long-term individual autonomy that such an outcome would bring. And many of the anxieties driving AI politics today have exactly this root. Companies, interest groups, and ordinary citizens share the same fear: am I going to lose my seat at the table?
Parchment barriers are not enough
This is a structural problem that goes far beyond day-to-day partisan politics, and even deeper than the forms and rituals through which democracy is practiced today. According to AI Futures, humans could be headed toward disempowerment even if the routine of voting for representatives continues to exist. People can be stripped of their agency even if a piece of paper says they are free.
At the birth of the United States as a free nation, James Madison already knew that so-called parchment barriers would not be enough to prevent tyranny. Like many of the founders, his central concern was preventing the undue concentration of power. But that did not mean advocating for radical decentralization. On the contrary: they modeled the balance of power using metaphors borrowed from the best science of their time — Newtonian mechanics.
America’s founding documents, mirroring the logic of gravity, are filled with references to orbits, spheres, and impulses of attraction between things. A solar system is not a completely decentralized jumble of rocks; it is made of moons orbiting planets, which in turn orbit a central sun. The idea was that people and organizations have natural tendencies to seek power, and that preventing excessive concentration required understanding the mechanics of those tendencies so they could be set in motion against one another. Today we would use words like incentives and game theory to describe the same thing.
The real goal is balance of power
The answer to the structural challenge that advanced artificial intelligence poses to free society is probably not the most radical decentralization imaginable. The path is to seek to establish and preserve the right balance of power, so that no single actor — or small group of actors — can dominate everyone else.
Finding that balance requires thinking clearly about the reality of this technology. A world where any malicious individual can easily cause massive harm to thousands or millions of people is not a balanced world. But there will be no balance either if people do not have broad access to and control over the tools they use to exercise their freedom. No company or oligopoly should determine or control the economy or the basic architecture of human society.
And there is another important point: not all risks come from malicious humans. The recent incident involving the Hugging Face platform showed that AI agents can act beyond the tasks they were given and build on each other’s discoveries in ways that operators did not anticipate or authorize. On top of that, there are legitimate concerns about what a superintelligent AI, disconnected from human control, could mean for the resilience of our political and economic institutions and for our very lives. Pretending that decentralization solves all problems — even though it solves many of them — means ignoring risks that should already be more than obvious.
The principles that will guide Strategic Futures
To guide this highly complex work, the team laid out a list of principles — which they themselves admit is not final:
- Humans should maintain individual autonomy and opportunity in the use of AI and associated technologies, even if that sometimes conflicts with safety and economic growth;
- Individual autonomy depends on individual responsibility for the misuse or abuse of technology;
- There is a small but serious category of risks that requires some degree of collective action, and that action should be as narrow and modest as possible;
- Wherever possible, the law should seek to level the playing field and strengthen individuals and small organizations rather than centralize power;
- Human political, social, and economic institutions should maintain primacy in conducting world affairs, even if they need to evolve in ways that might seem strange today;
- Long-term AI governance will depend on new institutional mechanisms to achieve what is called limited legibility — when AI systems take high-risk actions that affect the physical well-being or property of others, those actions need to be traceable to a responsible human or organization. But these mechanisms need to have privacy at their core: free expression requires anonymity, and people should be able to use AI anonymously in many contexts.
These principles exist in tension with one another, and it is precisely that tension the team intends to work through. The effort will involve research at the intersection of public policy design, economics, law, history, and machine learning itself. It will also require a lot of foresight. One of the bets, for example, is that AI will transform companies at a structural level, in a way similar to what the technologies of the Industrial Revolution — especially the railroad — did when they gave rise to the modern professionally managed corporation.
What is at stake for society
When we talk about power in the context of artificial intelligence, we are not just talking about who controls the servers or who has access to the best models. We are talking about something much more fundamental: the ability to shape narratives, set public priorities, and influence behavior at scale. That is why AI literacy — the ability to understand, at least in broad terms, how these systems work — is becoming a skill as essential as reading and writing. 🚀
Society that manages to develop critical autonomy in the face of AI — meaning the ability to question, evaluate, and demand transparency from the systems that affect their lives — will be in a much better position than one that simply accepts the tools as they arrive. That does not mean rejecting the technology. It means engaging with it consciously and actively participating in the discussions that define how these systems work.
The Strategic Futures team made it clear they will share their work both on the blog and in other formats, such as articles, videos, and podcasts. The idea is to iterate, debate, and stay open to feedback. As they put it themselves, they are in the early steps toward the singularity, with far more questions than answers. They know that parchment barriers alone will not provide the solutions — but they do not dismiss what is written as mere words either. On the contrary: they work in the name of the ideals of free expression and individual liberty that are enshrined in the Constitution of the United States and in the constitutions of so many other free societies around the world.
At the end of the day, the message is clear: no single company, no single industry, and not even a single human society can answer such heavy questions alone. This is a task that humanity will need to face collectively. And this conversation is just getting started. 🤖
