Views: 0
I must admit I don’t know. And, in all likelihood, no one does. But there is something that worries me: an increasing number of experts working directly in artificial intelligence are warning of extraordinarily serious risks if the development of this technology continues without sufficient safeguards.
Does this mean an artificial intelligence could wipe out humanity within ten years? We cannot say that for certain. It would be irresponsible to turn a possibility into a certainty. Yet, it also seems irresponsible to me to do the exact opposite: ignore the warning simply because it sounds too apocalyptic.
In 2024, a survey of 2,778 artificial intelligence researchers found that the experts' median estimate placed the probability of future AI advances causing human extinction—or an outcome of comparable severity—at 5%. Opinions were deeply divided, but the figure is striking enough that it cannot simply be dismissed with a smile.
And other events have occurred that should make us pause.
In recent years, top-tier researchers have left some of the leading AI companies, publicly voicing concerns about safety. One of the most well-known cases was that of Jan Leike, head of safety research at OpenAI, who stated in 2024 that safety culture and processes were losing ground to product development.
More recently, in September 2026, researcher Jacob Coxon left Anthropic and publicly raised the possibility that increasingly autonomous AI systems could become an extraordinarily grave threat within this decade. His claims are debated and do not constitute a proven scientific prediction, but they are part of a discussion that can no longer be considered fringe.
For me, this is the fundamental question. I do not have sufficient expertise to determine whether those predictions are correct. Yet I do not need them to understand that when people with deep knowledge of this technology voice concerns of such magnitude, the reasonable course of action is to investigate the issue with the utmost seriousness.
One need not believe the apocalypse is imminent. What is required is preparation for the possibility—however small—that it might occur.
In fact, the problem is not limited to a hypothetical AI that might one day decide to destroy us. There are far more immediate risks: cyberattacks, information manipulation, fraud, the use of AI to develop weapons, a loss of control over autonomous systems, and massive economic and social transformations. The International Report on AI Safety, backed by 30 countries and international organizations, points out precisely that significant uncertainties remain regarding future capabilities and the actual effectiveness of safety measures.
What surprises me is the lack of urgency perceived by society. We live surrounded by fake news, exaggerations, wars, catastrophes, and constant controversy. Perhaps that is why we have developed a sort of immunity to warnings. Yet there is a world of difference between being skeptical of a sensationalist news story and ignoring a technological risk that could have enormous consequences.
Governments should be seriously analyzing these scenarios, coordinating internationally, and establishing control and safety mechanisms before the technology reaches capabilities that become difficult to manage.
And we, as citizens, should pay closer attention. Perhaps in ten years we will discover that all these fears were exaggerated. Or perhaps we will find that we were too trusting. That is precisely why we must act now.
For if there is a reasonable possibility that we are creating a technology whose future behavior we cannot fully control, waiting for absolute certainty could be the worst possible strategy.
Artificial intelligence could become one of the greatest tools for progress in history. Yet precisely because of its immense power, it also warrants something we may not have sufficiently devoted to it until now: prudence, vigilance, and responsibility.
EN 
ES