Who Will Dictate the Rules of the Game?

Recently, Anthropic's research team published an article arguing that democracies must lead the development and deployment of advanced AI. Their argument is clear: the political system in which the most powerful AI systems are created will determine global norms, values, safety standards, and even military and surveillance uses. From my perspective, they are not wrong. Every tool, or weapon, always has a dual nature: it can be an engine of progress or an instrument of destruction. That is the eternal duality of technology and the price we must pay for progress.

It is clear that Anthropic frames this scenario from a United States perspective, positioning the U.S. and its allies against autocratic powers such as China. As a company based in San Francisco, that position is logical and expected. However, although I largely agree with what they express in their paper, I would like to reflect on the implications of this leadership and on what place awaits old Europe within it.

Anthropic presents two possible scenarios and argues that the only safe path is for the U.S. to maintain a clear competitive advantage so it can exercise real influence over the global standard. And this is where I ask myself: is it really the U.S. government that holds the power in this equation? Do we want to leave a technology potentially more dangerous than nuclear energy in American hands?

Regarding the first question, every private company has a primary objective, regardless of whatever marketing varnish is applied to it: to generate profits and accumulate power in order to keep growing. I do not consider that goal perverse in itself. In fact, I consider it a laudable goal in the times we live in: it is legitimate to want to be the best, to lead a market, and to reward those who trust the project, that is, its shareholders. The risks, however, are not in the goal itself, but in the path chosen to achieve it.

History is full of companies that have knowingly twisted the law or deliberately broken it in order to maximize profits. The famous 1998 lawsuit against the major U.S. tobacco companies is one example, where it was shown that companies had deliberately hidden, for decades, scientific evidence about nicotine addiction and the carcinogenic effects of tobacco. They did so for a single purpose: to protect their profit margins at the expense of public health.

If we transfer that precedent to artificial intelligence, the dilemma is obvious. If the companies developing the world's most powerful AI prioritize quarterly growth and competitive advantage above safety or the democratic values they claim to defend, can we blindly trust that U.S. leadership will be "more ethical" simply because it takes place within a democracy?

By this I do not mean that the development of these models should be exclusively public, because the incentives of the politicians who run states will be to use them for their own benefit. Think, for example, of public television channels and how politicians on one side or another distort them. Perhaps the answer lies in the separation of powers. Of course, I do not have a clear answer to this dilemma, but that does not make it any less worthy of being raised.

This reflection leads us to a second question that all Europeans should ask ourselves. Europe, whose countries are systematically ranked among the freest and most democratic in the world, is at a crossroads: artificial intelligence can become the great defender of our values or their greatest threat. While in recent times European citizens have gained renewed awareness of our physical security and territorial sovereignty, I still detect a gap in our collective perception: we do not fully understand that the digital dimension will increasingly become the master pillar of our global security.

Disinformation campaigns, hacks, electoral interference... Faced with this landscape, an almost visceral question arises: why not strike back with the same currency? The answer is as simple as it is unsettling: because, basically, we cannot even deter them.

Europe has defined itself as the "moral referee" of the digital world. Our laws, such as the recent AI Act, are designed to protect citizens, which is all well and good. However, this creates an enormous strategic vulnerability. We must ask ourselves: who really holds power, the country that is the only one capable of offering strategic services, or the country that is forced to buy them?

The reality is uncomfortable. No matter how many multimillion-euro fines the second tries to impose, the first knows it has the winning card: dependency. The provider understands that it only has to threaten to "turn off the technological tap" for the regulator, faced with the abyss of falling behind or seeing its economy paralyzed, to be forced to relax its laws. It is an implicit form of blackmail in which innovation becomes a weapon of political negotiation.

In conclusion, Anthropic's post is right about one thing: the place where AI is built will dictate the rules of the future. But for Europe, it is not enough to hope that the "democratic leader", the U.S., will protect us. We need to stop being merely the regulator of the game and become a player with our own capacity. Because in the age of advanced AI, freedom is not guaranteed only by laws written on paper, but by sovereignty over the code that executes them.

Are we still in time to wake up? Or will we remain simple consumers of a technology whose values have been programmed thousands of kilometers away?

References

• Anthropic (May, 2026). 2028: Two scenarios for global AI leadership.