An Assessment of the AI Gap Between China and the United States

The gap between China and the United States in the field of Artificial Intelligence (AI) has been the focus of attention at home and abroad. From different angles and dimensions, we can draw different conclusions. Here are some summaries based on recent relevant reports and analysis.

Comparison of Technical Papers and Patents

According to an article in The Economist, China's share of AI papers in 2019 exceeded that of the U.S. In 2021, 26% of global AI papers came from China, while the share from the U.S. was only 17%. From this perspective, China appears to be leading the way in terms of research output in AI1.

Comparison of base models and applications

However, in the area of "base models" that give generative AI its intelligence, the US has a clear advantage. chatGPT and the advanced model behind it (the latest version is GPT-4) were developed by US startup OpenAI. Other U.S. companies have powerful systems of their own, ranging from smaller companies such as Anthropic or StabilityAI to tech giants such as Google, Meta, and Microsoft (which owns a portion of OpenAI)1.

Comparison of learning capabilities and application potential

According to Zhou Hongyi, founder of 360, the gap between China and the U.S. in AI lies mainly in "determining the direction of technology". Once the direction is determined, China's advantage is the ability to learn quickly, and this gap should be able to catch up within one or two years. Zhou Hongyi also mentioned that 2024 may become China in the field of AI "application of the year", and believe that this year, the big model will be in many vertical areas of the enterprise is very promising 2.

Hardware and Talent Comparison

The second reason for China's mediocre performance in generative AI has to do with hardware. The U.S. has imposed China-facing export controls on a number of key technologies in AI, including microprocessors used in cloud computing data centers (where the underlying models learn) and chip fabrication tools that would allow China to make such semiconductors on its own. This has dealt a blow to China's big-model R&D.1 In addition, talent is a shortcoming in China's AI development. For now, the U.S. remains highly attractive to global tech talent: two-thirds of U.S. AI experts publishing papers in journals are foreign-born. in 2019, engineers of Chinese descent accounted for 27 percent of this top group. Many Chinese AI researchers have studied or worked in the U.S. and then returned home with specialized knowledge. However, the New Crown epidemic and increased tensions between the U.S. and China have led to a decline in this group1.

Conclusion

In summary, the gap between China and the United States in AI is not simply measurable in a few years, but is a comprehensive gap involving technology, talent, hardware, and many other aspects. Although China has made significant progress in some areas, there are still major challenges in basic technologies and key areas. However, with strong support from the Chinese government and continuous innovation from the business community, China's AI industry still has great potential and potential for development.