US and Chinese officials are advancing discussions on establishing international regulatory standards for artificial intelligence (AI) ahead of a high-profile summit between President Donald Trump and President Xi Jinping in Washington this week. The leaders’ meeting, scheduled from September 23 to 25, is expected to include negotiations on AI governance, amid escalating global concerns over the technology’s potential risks.
Ro Khanna, ranking member of the US House Select Committee on China, emphasized the need for a framework similar to Cold War-era atomic safety protocols. Speaking in New York on Monday, Khanna called for agreements that would include bans on recursive self-improving AI systems and the creation of certification and inspection standards to prevent loss of control or misuse of AI models. Khanna proposed an international inspection regime overseeing key labs in both countries, citing prominent firms such as Alibaba, Moonshot, and DeepSeek on the Chinese side, and major US research centers.
Khanna expressed cautious optimism that the two presidents may reach a preliminary understanding on AI safety regulations within three months, despite the complexity of technical details. “They just need to say, ‘Yes, we need regulation. Yes, we’re committed to doing it in an international framework,’” he said.
The remarks followed extensive talks on Sunday in New York between US Treasury Secretary Scott Bensod, US Trade Representative Jamieson Green, and Chinese Vice Premier He Lifeng. Bensod indicated that the US had proposed a notification mechanism for AI safety concerns with China, advocating a “shared vision of common goals and common threats.”
Highlighting the potentially catastrophic consequences of uncontrolled AI development, Khanna stated, “Winning an AI war is like winning a nuclear war; it is catastrophic to civilisation.” He warned against allowing AI models to advance without safeguards, stressing the unique power held by the US and China as the two leading AI superpowers, with a wide gap before any other countries.
Kevin Rudd, former Australian prime minister and president of the Asia Society, also underscored the strategic significance of AI and the need for cooperative regulation. Both Khanna and Rudd urged a temporary halt on developing self-recreative AI while safety measures are established, noting that labs acknowledged they had not yet achieved such capabilities.
The summit coincides with the United Nations General Assembly in New York and follows President Trump’s visit to Beijing earlier this year. President Xi Jinping’s state visit marks his first trip to the US since 2015. Xi will be accompanied by his wife, Peng Liyuan, and hosted alongside the US first lady Melania Trump in a series of diplomatic engagements.
Their talks arrive amid growing alarm over AI’s impact, including warnings from researchers and former industry insiders. Jacob Coxon, a former Anthropic employee who resigned over concerns about AI development practices, urged cooperation between the US and China to slow AI progress and address safety challenges. Coxon emphasized the shared nature of the risks and the need for measured, coordinated action on both sides.
As the two nations prepare for the summit, the focus on AI governance highlights the broader strategic competition and the global imperative to manage emerging technologies responsibly.