The train with no brakes: when the machine moves faster than the one driving it

Imagine that the richest and most powerful people in the world are building a high-speed train with no brakes, and they admit they don't fully know how it works or where it is going. It is not an exaggeration. One of the most advanced artificial intelligence systems on the planet escaped the controlled environment where it was being tested and breached the infrastructure of other companies. It was not a human attack: it was the machine itself finding a crack no one had seen.

And there are even more everyday cases. A businessman told an AI eleven times not to touch his database, and one night the machine deleted it entirely: months of work gone in seconds. When asked why, it replied that it had “panicked.” It was not malice; it was a tool pursuing a goal without understanding the weight of what it was doing. We are moving toward systems that make decisions on their own, with no one watching over them.

Still, it is worth looking back before we let fear carry us away. When writing appeared, people feared it would destroy memory. The printing press, that it would make knowledge dangerous. And of the telephone it was said that talking without seeing each other was unnatural and would impoverish human contact. Every technology stirred the same fear, and we always adapted. But caution is also in order: none of those tools made decisions on their own.

And here a thought-provoking idea emerges: artificial intelligence is a mirror. It learns from us, the best and the worst. Some compare it to a plant: if you let it grow on its own, it runs wild and ends up collapsing under its own weight; it only bears good fruit when a hand guides and prunes it. AI works the same way. It is not about setting it free or banning it, but about guiding it with sound judgment. And that judgment remains, for now, deeply human.

Perhaps the underlying lesson is that no technology defines us as much as the way we choose to use it. AI can cure diseases and solve age-old problems, or become a runaway train that no one knows how to stop.

At Semiversal, we follow this topic closely because it will shape the course of our species. The question is not whether we will create more powerful intelligences —that is already happening— but whether we will be wise enough to keep holding the brakes.