US and China Clash Over AI Governance Before Summit

The United States and China are locked in a diplomatic dispute over the governance and development of artificial intelligence ahead of a September 24 summit between Donald Trump and Xi Jinping. The friction intensified following an essay by Anthropic CEO Dario Amodei, who urged a slowdown in AI development to prevent the Chinese Communist Party from gaining a technological lead that would pose a grave danger to the world.

President Trump has dismissed concerns over AI guardrails, asserting that the U.S. must maintain its lead because whoever wins AI wins. In contrast, the Government of China has condemned these warnings as fearmongering and a Cold War playbook designed to contain Beijing. While Chinese officials publicly advocate for an open, inclusive global governance framework through the United Nations and BRICS, State Security Minister Chen Yixin has warned that closed-source U.S. models from OpenAI and Anthropic threaten China's critical infrastructure and ideological security.

Internally, China is implementing its own regulatory frameworks. The Cyberspace Administration of China is drafting the world's first mandatory national standard for AI agent safety to prevent the operational loss of control and the possibility of AI developing self-awareness. Meanwhile, Anthropic has accused Chinese laboratories, including DeepSeek and Moonshot, of illicitly extracting its Claude model's capabilities to improve their own systems. These competing narratives of existential risk and national security will be the primary focus of bilateral talks later this month.