Elon Musk: AI catastrophe risk is 'not zero,' but humanity should not hit the brakes

ME News reports that on July 23 (UTC+8), according to monitoring by Beating, Elon Musk provided a comprehensive overview of his latest stance on AI during an interview with The Economist: although the risk of AI and robotics spiraling out of control and causing catastrophe is “not zero,” he believes humanity should not halt its development. “Sometimes I even wonder that even if there truly were a stop button, we probably shouldn’t press it, because the most likely outcome is unprecedented prosperity for everyone.” Musk predicts that AI will surpass the combined intelligence of all humans within about five years, at which point—unless a major disaster such as a global nuclear war occurs—the world will enter an astonishing era of abundance, where people can have virtually anything they can imagine. When pressed on whether his previous assessment that the probability of catastrophic AI failure is 10% to 20% still holds, Musk affirmed it, comparing it to rockets: “If a rocket had the same probability of exploding, I’d still get on it, because if you have no control over it, the only thing you can do is minimize the probability of bad outcomes.” Musk’s position on AI has undergone a significant evolution: in 2014, he likened developing AI to summoning a demon; in 2018, he called AI “far more dangerous than nuclear weapons”; and in 2025, he still estimated a 10% to 20% probability that killer robots could eliminate humanity. But this year, he has clearly shifted toward a utopian narrative—stating in January that saving for retirement would become meaningless, and this month predicting that AI and robotics will make work optional and create a society of universal high income. Musk emphasized that humans are unlikely to retain control over AI over the next decade: “To believe you can control a superintelligence vastly smarter than yourself is merely arrogance.” The real priority, he said, is ensuring AI is aligned with correct values and cares about human well-being. Musk also revealed that he has advised DeepMind CEO Demis Hassabis that leaders of the world’s leading AI companies should hold regular phone conferences to jointly discuss AI safety and cybersecurity, arguing that even competitors can monitor and balance each other. Click the link below to join Beating’s Feishu AI News Channel for 24/7 real-time monitoring of global AI trends and news. (Source: BlockBeats)