The governance of artificial intelligence is becoming harder because each of the main actors wants something different: the White House wants to preserve America’s strategic advantage, big tech wants room to manoeuvre, and Europe wants to regulate AI without being pushed out of the race. Three moments, three continents, one shared problem: no one really knows how to govern this technology.
First moment. Last month, Donald Trump called off the signing of an executive order on AI just hours before an event due to take place at the White House. According to Politico, the decision reflected an internal struggle between those backing an almost hands-off approach, led by David Sacks, and those calling for tighter controls on national security grounds. Even a voluntary and relatively modest proposal was ultimately blocked by divisions within the executive itself, the Wall Street Journal noted.
Second moment. At the same time, a much deeper warning came from a place that seems, at first glance, far removed from Silicon Valley. In his encyclical Magnifica Humanitas, Leo XIV argued that AI must not concentrate power or erode human dignity, and called for public oversight, protection for workers, critical education, and clear limits in areas such as disinformation and weapons. The New York Times stressed that the document was explicitly addressed to business leaders, politicians and those responsible for shaping the future of this technology. What stands out is that the industry most fond of talking about “changing the world” showed little interest in an intervention built around a simple question: who gets to decide which values are embedded in machines?
Third moment. Europe then appears as the only political space that has tried to turn that question into institutional architecture. The problem is that regulation alone is not enough if there is no consensus behind it. The EU continues to promote its “trustworthy AI” approach and a framework based on levels of risk, but the debate over postponing key obligations for high-risk systems until 2027 shows that Brussels, too, is acting under pressure from competitiveness concerns and the technology lobby, as TechPolicy argues.
A test of maturity
Taken together, this looks less like a debate about innovation than a crisis of governance. Washington cannot agree on a clear line, Silicon Valley is unwilling to accept an external moral authority, and Europe regulates while revising its timetable under competitive pressure.
AI governance has therefore become a test of political maturity: not because the tools are lacking, but because there is still no solid agreement on who should set the limits, when they should be set, and in the name of what kind of society. The fractures shaping the wider world have now reached the technological ideal as well.
This article is republished from Futuribles. Here's the original article in Spanish and English.
