What's Happening?
China has proposed new artificial intelligence (AI) export curbs designed to promote a business model that keeps valuable technology within the country while selling access to it abroad, as rivalry with the U.S. intensifies. The draft rules, released
on October 16, differentiate between foreign customers using Chinese AI services and foreign companies seeking to obtain the underlying models, data, or expertise to operate or reproduce them independently. These regulations would generally permit overseas users to access Chinese AI via websites and cloud platforms, but would subject certain transfers of core technology or overseas deployments to government approval. The draft aims to add 23 AI-related technologies to a list banning or restricting their export, including advanced models, agents that perform tasks with limited human direction, and technology used for chip design and managing large computing facilities. A temporary requirement for registering certain transfers took effect immediately.
Why It's Important?
These proposed export curbs are a significant development in the U.S.-China tech competition, mirroring U.S. restrictions on China's access to advanced technology and investment. As Yang Jingru, a partner at law firm Jincheng Tongda & Neal, noted, 'The U.S. controls the inflow of computing power, hardware and capital, while China is beginning to systematically manage the outflow of models and technological assets.' This strategy aims to safeguard China's 'AI crown jewels' and foster indigenous innovation, potentially accelerating its self-sufficiency in critical technological areas. For U.S. businesses and global tech companies, these restrictions could complicate international partnerships, impact how advanced models are run on overseas data centers, and influence foreign investment in Chinese AI assets. The broader regulatory oversight adds to the uncertain macro-environment for tech giants and startups, particularly those grappling with the meltdown of Chinese AI and semiconductor stocks.
What's Next?
China is expected to adopt these proposals by the end of the year, with implementation guidelines to follow. Companies are seeking clarification on various aspects, including whether they can train or run their most advanced models on foreign-owned overseas servers and the specific tests used to determine which models are covered by the restrictions. The draft specifies three tests: computing work needed to train a model (threshold of 10^26 calculations), model size (at least 2 trillion parameters), or ranking among the world's top 10 in certain performance assessments. These rules could extend beyond model releases to corporate deals, with registration requirements already in effect for transfers that place Chinese technology under foreign ownership or control. This could lead to a more fragmented global AI ecosystem, with potential implications for international collaboration and the development of AI technologies.
Beyond the Headlines
The new export curbs highlight a deeper strategic objective for China: to maintain technological advantage and national security in the face of escalating geopolitical tensions. By controlling the outflow of its AI innovations, China aims to prevent its valuable intellectual property from being leveraged by rivals, while simultaneously fostering a domestic ecosystem that can thrive independently. This approach could lead to a 'decoupling' of AI development, with distinct technological stacks emerging in the U.S. and China. The ethical implications include questions about the free flow of scientific knowledge and the potential for a bifurcated global technological landscape, which could hinder universal standards and collaborative solutions to global challenges. The long-term impact could be a more resilient but less interconnected global technology sector, with significant consequences for innovation, economic growth, and international relations.
