Among the most famous sayings of Imam Ali ibn Abi Talib, may God be pleased with him: "If you are in retreat and death is advancing, how swift the meeting must be."
I recalled this wise saying amid the clamor that accompanied the words of Jacob Coxon, the British researcher who resigned from the artificial-intelligence company Anthropic- words I do not think he exaggerated.
Coxon is anxious about the absence of legislation binding artificial intelligence companies to a single set of safety and security standards. He warned that this advanced technology may, within a few years, put an end to the human race, particularly once its capacities develop to the point of surpassing the intelligence and abilities of its own developers. Artificial intelligence thus slips its tether and begins a phase of taking its own decisions without reference to the human being, and indeed against his will.
Coxon, it should be said, is no exceptional case among the minds that have known AI from the inside and then come to realize that man is dealing with a technological monster that will threaten humanity's fate. The alarm bell has been ringing since this technology began to threaten a great many things in our lives: employment, politics - elections and governance specifically - the integrity of the judiciary, and the curricula and methods of education.
One of the pioneers of artificial intelligence, the British-Canadian professor Geoffrey Hinton, had already voiced his apprehension at the start of 2023 about the speed of its development. In the spring of that year, he resigned from his senior position at Google in order to devote himself to speaking about the dangers of AI, without failing to note that a part of him regrets his work in the field.
Among the aspects Hinton addressed in media appearances was the danger of this advanced technology surpassing the human being's cognitive capacity to keep up. He said the most alarming thing about it was the chatbot, the program that simulates human speech in voice and in text. Over the following two years, the man's pessimism deepened regarding the misuse of AI and the escape of this technology from any controls. By the end of 2024 he was indeed putting the probability of it causing the end of the human race at between 10 and 20 percent within three decades alone.
But leaving aside the ivory tower in which scientific thinking does or does not reside, there are things happening on the ground.
Coxon's words provoked the global trillionaire Elon Musk, one of the technology oligarchs most deeply tied to political movements, and perhaps the most prominent financier of far-right parties across the Western democracies. Musk responded to Coxon forcefully, which drew a sharper response from his ideological opposite, the left-wing Democratic senator Bernie Sanders, on X.
Musk had counted Coxon's words a smear, or a trap. Sanders replied in the manner of condemning a man out of his own mouth: that Musk himself had said in 2014 that "with artificial intelligence we are summoning the demon."
Then in 2018 he said: "Mark my words, AI is far more dangerous than nukes. Far. So why do we have no regulatory oversight? This is insane."
And even in 2025 he said the earth faces a threat of "annihilation by killer robots" of between 10 and 20 percent. Sanders concluded by asking Musk what had changed now, and whether the shift in his position had anything to do with his investing billions of dollars in AI and planning to develop ten million robots a year.
Setting aside personal political positions, I believe the alarm gripping informed circles around the world is a positive development, in light of scientists' fear of the lethal dangers of AI slipping its restraints. Here my memory returns to our forebears' use, since remote ages, of the expression "releasing the marid or the jinni from the bottle."
This is indeed what is happening now. But the difference is that the marid, or the jinni, needed someone to release it, by rubbing the bottle, whereas the dangers of today and tomorrow suggest it no longer needs anyone to release it. It will act as it pleases.
Add to this that modern and future technology, AI now included, has penetrated every aspect of our lives. We are surrounded today by algorithms of every type and specialization, to the point that those who own the means of knowledge, through eavesdropping, espionage, imaging, aggregation, or analysis, know more about each one of us than that person knows about himself.
As far as I am aware, there are companies that moved beyond identifying people through facial scanning some time ago and have come close to - if they have not genuinely arrived at - scanning intentions and predicting reactions in advance.
At this point, what value remains in opinion polls and their analysis, if manipulating them is easier than drinking water? What hope of change remains, given a strategic alliance between right-wing radicalism, the lords of capital, and the magnates of advanced technology?
In confirmation of this, and in addition to Musk's role in supporting extremist forces across Europe, it was revealed only yesterday that two cryptocurrency billionaires had donated, in equal shares, £72 million to the British Reform party.
