The growing chorus for caution on artificial intelligence can no longer be dismissed as the anxiety of technology pessimists. Nor should it be interpreted as a call for the United States or China to surrender their race for technological leadership. The message from several of the world’s most influential technology leaders is much simpler: do not allow competition to become so rigid that safety is sacrificed at the altar of speed. That is a message both US President Donald Trump and Chinese President Xi Jinping would do well to heed. Trump has made his position unmistakably clear. He has said the United States is ahead of China in AI and wants to keep that advantage, arguing that whoever wins AI will occupy an extraordinarily powerful position. There is considerable evidence that the US still leads China in several important areas of frontier AI, including research and model capabilities. At the same time, China is closing the gap rapidly, making the competition far from settled. But if Trump believes America is indeed comfortably ahead, that should be precisely the reason to listen more carefully to the warnings coming from within America’s own technology establishment. Bill Gates has warned that AI is advancing faster than governments and societies can prepare for its consequences, including disruption to both white- and blue-collar employment and new cyber and biological risks. Elon Musk and other technology leaders have raised concerns about increasingly autonomous and potentially uncontrollable AI systems. Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman have also supported calls for greater restraint around frontier AI development.
These warnings cannot simply be labelled anti-American or anti-innovation. Nor does Mark Zuckerberg’s position fit neatly into the “slow down” camp. He has argued that excessive concentration of superintelligent AI in the hands of a few governments or companies could itself become a major danger. Meta’s current position stresses broad access, individual empowerment, and safety measures while continuing rapid development. That divergence among technology leaders is important. They disagree on the solution, but they increasingly agree that the stakes are enormous. China, too, should not dismiss these concerns as merely another American attempt to contain its technological rise. Beijing has rejected the US “threat narrative” surrounding Chinese AI development, while Chinese officials have simultaneously acknowledged serious risks arising from artificial intelligence. The sensible course for both Washington and Beijing is therefore neither an uncontrolled race nor an unrealistic global freeze. Compete, but communicate. Innovate, but verify. Lead, but build safeguards alongside capability. The US and China do not have to agree on every aspect of AI governance. They do, however, have a shared interest in preventing an advanced AI system from causing consequences neither side intended. Trump’s confidence in America’s technological lead should encourage him to negotiate from strength rather than fear. Xi’s determination to make China technologically self-reliant should similarly not prevent Beijing from engaging with legitimate international concerns about AI safety. The irony would be profound if two superpowers racing to outwit each other ended up creating a technology neither could fully control. AI may well become one of humanity’s greatest achievements. But the race to reach the summit should not become a race in which the competitors forget to ask whether the mountain is safe to climb. Trump and Xi have an opportunity to ensure that technological rivalry remains a contest for human advancement—not a contest in which caution itself becomes the first casualty.
