On September 26, 1983, as the world, caught in the grip of the Cold War, labored under the weight of mutually assured destruction, an act of disobedience saved humanity. That night, the sophisticated Soviet detection system signaled the launch of five American intercontinental ballistic missiles toward the U.S.S.R. Soviet officer Stanislav Petrov, exercising his free will unconstrained by any computer program, decided to defy military protocol that demanded a massive retaliation. Instead of plunging the world into an unprecedented nuclear conflict, he chose to report a false alarm to his superiors. History would record that his decision not to trust the machine, and not to trigger the fatal sequence, was the wisest choice. What the computers interpreted as a missile launch turned out to be nothing more than sunlight reflecting off clouds, which the system had mistaken for infrared signatures. Petrov possessed an intuition that machines do not share: that a genuine American offensive would include hundreds of missiles, not merely five. Trusting his own reasoning rather than being guided by the machine proved to be the right approach that day. He exercised a capacity we must preserve in the face of machines: doubt. And perhaps in certain other moments of our era, marked by the omnipresence of artificial intelligence, doubt must be revalued. Our seconds, minutes, days and weeks are too often influenced by decisions that do not truly emanate from us. We risk falling into certain traps that could open in moments when critical thinking is needed, yet its use increasingly fades from the list of fundamental reflexes among many of our contemporaries. Like Ulysses, who had himself bound to a mast to escape the sirens' song, we too must cling to an anchor point that would allow us to hear the machines without succumbing to them. To doubt and keep our faculty for thought active. This may be our best defense against what artificial intelligence could cause. Millennia ago, Socrates made awareness of our ignorance the starting point of knowledge. Today, answers to our questions, obtained in a fraction of a second, deserve more than ever to be questioned for their truthfulness. Like Socrates in his time, who interrogated the rhetoricians, our relationship with artificial intelligence must be a conversation, not a surrender. For even if current computer tools can help enrich our thinking, we must neither neglect nor abandon our role as guardians. Fenitra Ratefiarivony