Indian-American Congressman Ro Khanna has called for an enforceable US-China treaty designed to ban artificial intelligence systems capable of improving themselves without human training, warning that such technologies could cause humanity to lose control. Representing Silicon Valley as a California Democrat, Khanna emphasized that the two global powers should establish a specialized working group consisting of American and Chinese technologists to formulate common safeguards for advanced AI systems. Speaking on NBC’s Meet the Press, Khanna noted that banning recursive self-improving AI should be a universally shared objective, pointing out that China, as a tightly controlled society, faces an even greater risk of losing oversight over autonomous models.
Monitoring Frontier Labs and Verification Frameworks:
Khanna argued that the United States is well-positioned to lead negotiations toward an enforceable and verifiable treaty that includes strict monitoring of the world’s most advanced frontier AI laboratories. This proposal follows an agreement reached between U.S. President Donald Trump and Chinese President Xi Jinping to initiate formal bilateral discussions addressing the risks associated with artificial intelligence. While welcoming this initial dialogue as a positive first step, Khanna is actively backing legislative measures introduced by Senator Bernie Sanders and Representative Greg Casar. This legislation aims to temporarily pause the development of advanced recursive AI systems until robust regulatory safeguards can be successfully implemented.
