The technology war between the United States and China over artificial intelligence (AI) hegemony has entered a new phase. As the U.S. unveiled an AI blueprint featuring 'America-centered AI alliance prioritization,' China quickly countered by emphasizing 'global cooperation.' This reflects the stark divergence in positions between the U.S., which has spent decades pouring enormous capital and talent to achieve technological superiority, and China, which must rapidly catch up. One side aims to reinforce alliances secured with overwhelming technological power while raising technological barriers, while the other is looking to build new coalitions to expand its influence, indicating an escalation in the rivalry between the two countries regarding leadership in the AI era.
The Financial Times reported on the 28th (local time) that China showcased its AI technology at the 'World Artificial Intelligence Conference (WAIC)' in Shanghai from the 26th to the 28th. More than 800 Chinese corporations participated in the event, where they exhibited various robots utilizing AI, including robots boxing with the latest domestic AI models, robots peeling eggs, and robots playing Mahjong.
Li Qiang, the Chinese Premier, proposed establishing a 'global AI cooperation organization' at the event. He noted, 'I oppose creating an exclusive 'small group' like some countries do' and expressed willingness to share AI development experiences and technologies with Global South countries. The Global South refers to over 120 developing countries primarily located in the southern hemisphere or the low latitudes of the northern hemisphere.
The 'small group' mentioned by Premier Li refers to the AI ecosystem centered around the United States. On the 23rd, just before the WAIC opening, U.S. President Donald Trump announced the 'AI Action Plan' and signed a related executive order. This plan is aimed at reinforcing the U.S.-led AI ecosystem by providing support at the government level for regulatory relaxation for AI corporations, infrastructure expansion such as data centers, and technology exports focused on allied countries. At the signing, President Trump made it clear that 'leading the world in the AI field at any cost will be our policy,' highlighting the principle of American primacy.
The U.S. AI Action Plan operates around closed (Proprietary) AI models such as OpenAI's GPT series and Google DeepMind Gemini, showcasing some of the world's leading technologies. Closed models do not publicly disclose the source code, which is essentially the 'AI recipe,' and manage it strictly to prevent technological leaks.
In contrast, China is using the guise of global cooperation to expand its influence as a latecomer breaking through constraints. According to Xinhua News, China already possesses more than 40% of the 3,755 AI models known worldwide, with 1,509 being developed under an open-source method. Open-source refers to a method of publicly sharing source code free of charge, allowing anyone to freely utilize and develop the technology. Representative Chinese open-source AI models include DeepSeek, which astonished the world with its performance despite semiconductor export controls, and Alibaba Qwen. The U.S. foreign policy journal Foreign Affairs has described this as 'China's calculated strategy to secure influence against U.S. AI technology control.'
Experts evaluated that the fundamental goals pursued by the two countries in the world of AI differ beyond a simple open or closed dichotomy. An analysis of the AI Action Plan signed by Trump shows that the U.S. aims to 'counter Chinese influence in international governance bodies.' The report pointed out that 'even international organizations such as the United Nations (UN) are being influenced by Chinese corporations attempting to lead AI surveillance standards, including facial recognition.' It further stated that 'we will lead the AI world with our allies, reflecting American values in AI governance.'
Perspectives on which countries to partner with are sharply divided. The Global South mentioned by China is broadly situated in Asia, Latin America, the Middle East, and Africa, centered on the low latitudes of the northern hemisphere and the southern hemisphere. These countries have begun to voice their independent positions following recent experiences with the COVID-19 pandemic, U.S.-China hegemonic competition, and climate change. In particular, they seek to maintain a strategic balance between the traditional ally group of the U.S., Europe, Korea, and Japan for urgent issues requiring international cooperation, and the coalition of China and Russia. Some argue that China is framing the narrative that 'countries with superior technologies like the U.S. are stepping on nations that are just beginning to grow.'
In contrast, the U.S. has made it clear that its principle is 'to export American AI only to allies and partners.' This strategy aims to bind allies into an American-centric ecosystem by supplying all U.S.-made AI technologies, from advanced semiconductors to the latest AI models and applications, in one go.
George Chen, a partner at consulting firm Asia Group, said, 'While China sticks to a multilateral approach centered on Belt and Road Initiative participants, the U.S. will likely seek to consolidate its own camp by rallying traditional allies like Japan and Australia.'
Philosophical differences regarding the role of government were also revealed. The U.S. AI plan identifies 'removing red tape' as its first task. The focus is on eliminating regulations that hinder innovation in the AI-related private sector, including startups. In contrast, China, like WAIC, takes a 'state-led' approach, with the government taking the lead in planning and directing the AI industry.