How many predictions about the end of the world can we tolerate before we start questioning their meaning and reliability?
Since ancient times, in many human cultures, figures such as holy men, prophets, and mystics have sought to interpret the course of history and the fate of humanity. In societies where survival depended on hunting or farming, these individuals often served as spiritual guides: they observed natural and celestial cycles, offered answers to the great existential questions, and helped communities make sense of events. In return, they received material support from those who produced goods and food.
With the rise of the major monotheistic religions, the perspective on time evolved in a distinctive way. The Abrahamic religions (Judaism, Christianity, and Islam) share a fundamentally linear conception of history: it has a beginning, unfolds according to a divine plan, and moves toward a definitive culmination. Unlike the cyclical visions found in other traditions—in which creation and destruction repeat indefinitely—these traditions propose a path oriented toward an ultimate end capable of giving full meaning to human existence and resolving the moral and theological tensions of life. This linearity imbues history with profound meaning, transforming it from a mere sequence of events into a narrative guided by a transcendent purpose.
Christianity, based on the Book of Revelation and the Gospels, views the end of time as the glorious return of Jesus Christ (Parousia), the Last Judgment, and the definitive victory of good over evil. In Islam, the end of the world is called Yawm al-Qiyamah (Day of Judgment) or Al-Sa’a (the Hour). It is preceded by minor and major signs, including the appearance of the false messiah (Al-Masih ad-Dajjal) and the return of Jesus (Isa), who will defeat evil to restore a period of peace before Allah’s final judgment.
From Apocalypses to Pessimistic Science Fiction
More recently, apocalyptic predictions have focused on a possible nuclear holocaust, overpopulation, and the depletion of resources, as well as climate disasters caused by human activity. Lately, computers and software have also entered the picture, first in Asimov’s science fiction novels, then with the drama of the turn of the millennium (Y2K), which was supposed to bring the world to a halt (and instead turned out to be a false alarm, albeit a costly one), and finally with today’s fears of Artificial Intelligence (AI) that will destroy us and which must, of course, be controlled and stopped by laws, taxes, and bureaucracy.
As for a possible nuclear holocaust, it could still happen, given the large arsenals still possessed by various countries (see below).
It should be emphasized, however, that, so far, nuclear weapons have brought a world war to an end, saving millions of lives and halting Japanese genocides in Asia (whatever certain analyses may say…). Furthermore, an honest and realistic reading of European history since 1945 should convince us that it is only thanks to the deterrence provided by Western nuclear arsenals that we have managed to avoid wars and invasions by Soviet tanks and troops.
But let’s turn to today’s topic: the claims—coming even from authoritative sources—that AI could, in the near future, destroy humanity. For example, Jacob Coxon—a former researcher at Anthropic—resigned to warn that the top AI labs are acting irresponsibly. Evan Hubinger, Anthropic’s head of alignment science, stated that there is a 10% probability of a human extinction caused by AI within the next 10 years. Another well-known expert at Anthropic—Nate Soares—has echoed this warning.
Even Bill Gates, without talking about the end of the world, has stated that the rapid and uncontrolled rise of artificial intelligence will trigger one of the most turbulent periods in human history, asserting in no uncertain terms that there is no plan to address the massive social upheaval that lies ahead. Let’s remember, however, that Bill Gates rode the wave of climate alarmism for decades, only to recently change his tune with rational and wise statements. Perhaps he’ll do the same with AI when it suits him.
The Political Response to Alarmism
There are many reactions against this new wave of alarmism. These reactions span a broad political spectrum. On the right, President Trump is at the forefront, dismissing warnings about the existential risks of artificial intelligence as “fake news” and a “sick conspiracy,” and arguing that future government oversight would stifle American innovation and give an advantage to hostile competitors like China.
On the left, however, we’re told that (1) exaggerating the threats posed by AI makes this software seem magically powerful, keeping investment bubbles afloat and impressing Wall Street; (2) calling for strict government oversight allows big tech companies to write rules that exclude smaller startups and competitors; and (3) portraying themselves as reluctant guardians against an unstoppable apocalypse helps executives at major AI companies avoid accountability for future mistakes and user complaints.
Furthermore, the fact that much of this alarmism comes from Anthropic itself is highly suspicious, given that the company is planning its initial public offering (IPO) for November 2026. Anthropic is seeking to raise up to $100 billion in the IPO, which would value the company at $2 trillion. It’s fair to wonder whether this alarmism from Anthropic is part of a rather hypocritical strategy to tout itself as the only AI company that promotes rigorous safety frameworks and exercises strict control over corporate power, practices, data, and military contracts. Pecunia non olet.
If doomsday scenarios distract from concrete problems…
In short, skeptics point out that panic over science-fiction-style catastrophes distracts attention from verifiable, everyday issues. These include (1) the bias of many users toward AI algorithm decisions, especially when they observe a strong left-wing political influence in the AI’s responses to questions about history and current events; (2) the massive energy consumption of data centers, which are indispensable for enabling AI to provide comprehensive and rapid answers to the questions it receives; (3) vulnerabilities in cybersecurity; and (4) the erosion of educational standards, especially regarding students’ use of AI to conduct research and write assignments and theses.
This last point—educational standards—is significant. For many years now, there has been an underground market—worth $1 billion a year—that allows college students in Europe and the United States to pay experts in the developing world (especially in Kenya) to do part of their academic work. For some, writing college essays as ghostwriters for students who lack the time or inclination to write them can be lucrative, especially in countries with high unemployment among young college graduates. Well, now AI is ruining this semi-legal market by offering low-cost computer tools that, in many ways, can replace the work done by African ghostwriters.
As usual, every revolution brings unexpected consequences, both positive and negative.
