China May Soon Be the Top AI Innovator in the World
WASHINGTON— China’s relentless drive and strategic investments in artificial intelligence (AI) indicate that it is only a matter of time before it matches or surpasses the United States’ capabilities in the field, according to a new report from the Information Technology and Innovation Foundation (ITIF), the leading think tank for science and technology policy.
“The narrative that China is merely a copier is false and outdated,” said Senior Policy Manager Hodan Omaar, who authored the report. “Now, it has developed an AI innovation ecosystem that rivals that of the United States, producing cutting-edge models that exceed those of its global peers. If this trend continues, the U.S. may soon find itself in the position of copying China.”
Despite extensive U.S. efforts to restrict China’s access to advanced technology through export controls, these measures have had limited success. On the contrary, they have been counterproductive, spurring China to advance its homegrown ecosystem, with companies circumventing restrictions by training in public clouds or using on-premises private cloud solutions like Huawei’s “AI-in-a-box” products.
Key findings in the report include:
- China leads in AI research publications and is competitive in generative AI, though U.S. publications have greater impact, with more citations and private-sector involvement.
- Tsinghua University is a key hub for China’s top AI start-ups, including "AI tigers" Zhipu AI, Baichuan AI, Moonshot AI, and MiniMax, founded by faculty and alumni.
- Chinese large language models are closing the performance gap with U.S. models, with some excelling in bilingual benchmarks.
- While private AI investment in China is lower than in the U.S., foreign investment, particularly from Saudi Arabia’s Aramco, is increasing in China’s generative AI sector.
- State-directed funds and financial aid effectively support promising firms in underinvested regions of China.
Chinese start-ups are influencing even top Western institutions. A recent controversy at Stanford University revealed that its AI model Llama 3-V closely resembles MiniCPM-Llama3-V 2.5 developed by Tsinghua University’s NLP Lab and ModelBest. ModelBest co-founder Liu Zhiyuan noted the significant gap between China’s generative AI models and Western projects like GPT-4 but emphasized that China has rapidly evolved into a key driver of AI technology innovation in just over a decade.
To enhance AI development and increase adoption, Omaar recommends U.S. policymakers:
- Stimulate private investment in AI research and development.
- Revitalize federal funding processes for AI.
- Avoid policies that undermine U.S. leadership while strengthening Chinese competitors.
- Develop a national data strategy to expand data availability for training models.
- Create a national roadmap for adoption.
- Prioritize rapid adoption within federal government agencies.
- Support digital transformation initiatives.
- Incentivize workforce training investments.
“The United States is the global leader in AI development but faces fierce competition from China,” Omaar concludes. “Focusing on stifling China's growth will not halt its progress long-term. Instead, maintaining or expanding its lead should be prioritized through comprehensive national strategies.”
