China defends DeepSeek after US ban of its AI from government systems

China responded to the U.S. ban on its AI startup DeepSeek by warning of retaliatory measures, highlighting the ongoing tension between the two superpowers. The Department of Homeland Security instructed federal civilian agencies to remove DeepSeek products and third-party services developed using DeepSeek from their IT systems.

George Chen, a digital practice leader at The Asia Group consultancy, noted the "depth of strategic distrust" between Beijing and Washington. The Homeland Security order, from late July, also prohibits civilian agencies from procuring DeepSeek products and services. DeepSeek's open-source models are accessible to countries like the United States, and are typically free to use.

Chinese officials, however, rejected the U.S. ban, vowing to protect the rights and interests of Chinese companies and opposing the overstretching of national security concepts. They criticized the U.S. move as disregarding the Xi-Trump consensus and the broader China-US economic ties. China and the U.S. have occasionally collaborated on AI governance, but remain divided on many issues.

DeepSeek did not comment on the ban. The two countries recently made progress on AI governance, including a proposal for a notification system on national security incidents, but the issue remains contentious. AI is expected to be discussed when Trump and Xi meet again in mid-November at the Asia-Pacific Economic Cooperation summit in Shenzhen, China.

U.S. restrictions continue to bar China from accessing advanced AI chips and fabrication machines, while Beijing claims that Chinese AI startups, including DeepSeek, have misappropriated capabilities from leading American AI models, a claim China has denied. Last week, China announced its own proposals to limit exports of certain advanced AI models.