Who steers AI: China’s industrial state vs America’s frontier builders?

Artificial intelligence (which I will hereinafter abbreviate as AI) is generating great debate, both because of the risks associated with the development of new technologies, and because of the future opportunities they potentially herald for economic development.

China’s leadership has increased its focus on AI over the past decade, and in many contexts, the question that occupies minds is whether the US or China will become the most advanced nation in AI development. I think it is wrong to view China’s focus on AI as part of a “war” against the US (what evidence is there of China’s “attack”?), or a competition between the two countries (when will we know who has “won”?). It is far more interesting to understand the differences between the US and Chinese approaches to AI, as these may prove decisive in shaping how AI is used in both countries and across the rest of the world.

But first, it is worth understanding what AI stands for. Most people in the West think of the technology as “artificial” intelligence. In China, the technology is referred to as “rengong zhineng (人工智能)”, which can be more literally translated as “human-made intelligence”. When I emphasise this difference, it may be seen as sinological pedantry, but I think that there are very significant differences of meaning implied by the two terms.

The systems that go by the name AI are created by humans — for example, developed by Coca-Cola-drinking, pizza-eating programmers in “high-tech companies”. The systems are controlled by algorithms, i.e. instructions or “recipes” designed by programmers. The systems are trained with digital data, which is mainly retrieved from the Internet — i.e. “facts” mixed with the kind of nonsense that humans produce.

In other words, human-made “intelligence” is tainted with all the political “values”, prejudices and other stupidities that ordinary people — including programmers — suffer from. Advocates and opponents of AI are happy to give these technologies an air of something that is autonomous and outside of human influence — like robots that want to take over the world!

My point of view is, on the contrary, that AI is controlled by people, capital and intentions; that is precisely why it is important to control this technology to the extent that is practically possible — from its birth, development, to its eventual death!

China’s development of AI went on for many years without much international attention. In fact, the first seeds of AI systems development in China can be traced back to the 1950s, when Chinese researchers were influenced by cybernetics, a term for information processing that was popular in the Soviet Union. The development was also supported by Qian Xuesen (钱学森), a leading scientist in aerodynamics, who wrote a textbook on Engineering Cybernetics during the four years that he experienced house arrest, before the Americans deported him to China in 1955.

However, cybernetics was criticised as a bourgeois science in the 1960s and 1970s, and it was difficult to conduct research and development (R&D) for AI. It was not until the R&D tasks for intelligent computers and robots under the 863 High-Tech Program approved by Deng Xiaoping in 1986 that state support for the development of AI became a high priority.

During the 1990s, the digitalisation of the Chinese economy began and with new digital services came the need to develop AI. The competition between the leading digital platforms, known as BAT (Baidu, Alibaba, Tencent) created a rapid development of AI for search engines, among other things, in the first decade after 2000. In addition, a number of companies emerged that developed AI for special services, e.g. SenseTime for facial recognition and iFlyTek for automatic translation.

Since then, things have progressed rapidly and in 2017, the government formulated a strategy for AI called “A New Generation Development Plan for Artificial Intelligence” which envisions three phases up to 2030, in which Chinese theories, technologies and applications of AI will become globally leading and China’s AI industry will be internationally competitive. This plan did not cause much stir abroad, but brought together a number of Chinese companies that had already made a name for themselves in AI to further develop the technology for important application areas.

A breakthrough came in January 2025, when a hedge-fund company from Zhejiang province launched DeepSeek, a large language model known as R1 that could compete with chatbots and other AI-based tools like OpenAI’s ChatGPT. For American journalists, this was a completely unexpected development, because they expected China to be far behind and had no chance of catching up with the US. It was compared to the “Sputnik shock” that the US experienced in 1961 when the Soviet Union sent astronaut Yuri Gagarin into space with a Sputnik rocket. 

An interesting aspect of this story is that DeepSeek’s language model was a kind of byproduct of the project initiated by Liang Wenfeng, the CEO of High-Flyer, a Hangzhou-based quantitative hedge fund that owns DeepSeek, to use AI to improve the foundation of the hedge fund’s investments.

The cost of developing the R1 model was apparently 27 times less than the investment in OpenAI’s GPT-01. Moreover, DeepSeek was launched as “open weight”, meaning that the model’s trained parameters (weights) are publicly available. This created a trend for other Chinese AI companies, where many leveraged elements of DeepSeek’s models and launched open models of their own.

However, DeepSeek is just one of many AI companies in China. There are a number of competitors for DeepSeek-V3, such as Qwen 2.5-Max developed by Alibaba and Doubao 1.5 Pro launched by ByteDance. China’s AI industry will reach more than 6,200 companies and a turnover of more than 1 trillion RMB by 2025, according to Li Lecheng, the minister of industry and information technology. Li also stressed that AI has been applied in more than 30% of large Chinese companies by 2025.

The government’s policy on AI has evolved since the first AI strategy was announced in 2017. On 27 August 2025, China’s State Council issued a new action plan called AI+, which formulated more concrete ambitions for the country’s national AI strategy in the coming years.

Thus, they aim to achieve 70% AI adoption in key sectors by 2027 and 90% by 2030, with a further vision of building a fully AI-powered economy by 2035. This is also related to the plans for digitalisation in China, where the 15th Five-Year Plan (2026-2030) promises support for digital platform companies to play a significant role in promoting innovation, increasing employment and improving international competitiveness.

Moreover, the Chinese government is not shy about regulating the industry when it comes to the security of the state or citizens. The security of Chinese AI was a major topic during the Third Plenum in July 2024, and was also discussed as a special topic in a study session of the Communist Party Politburo in April 2025. Formally, the sector is regulated by China’s Cyberspace Administration, which approves algorithms and generative AI. In May 2025, the State Council also issued a White Paper that described AI as a “double-edged sword” that needed additional supervision and evaluation systems. 

This has become especially relevant during the OpenClaw craze that has gripped Chinese society. The OpenClaw app, also known as Lobster (longxia 龙虾), is an autonomous AI agent developed by an Austrian programmer that can organise people’s email, calendar, etc. It became hugely popular when Tencent offered to install it on people’s mobile phones earlier this year. But the Chinese government has immediately launched a project to formulate standards and tests for credibility regarding the use of OpenClaw.

It is important to remember that China has been building a robotics industry for decades, and that the country’s five-year plan will focus on “embodied-AI”, which stands for the complete integration of virtual AI with hardware. Many have probably seen China’s humanoid robots perform incredible martial arts stunts during the Chinese New Year celebrations. Whatever one thinks of such daring feats, the enthusiasm that both top politicians and a rapidly growing group of companies in China have for robots and the automation of manufacturing and services shows that a huge market for embodied-AI services will emerge in China.

The dancing robots also reveal a significant aspect of China’s evolution towards AI, namely a focus on specialised models, chips, and applications of AI in core economic sectors. This is somewhat in contrast to the strategy followed by the Americans.

In America, there is great pride in the fact that American companies are leaders in the manufacture of advanced chips for training and utilising AI. There is no doubt that a company like Nvidia is a leader in the design of chips and the manufacture of hardware for AI — where sanctions for exporting these chips to China became a challenge for Chinese AI companies. Nor is there any doubt that companies like OpenAI, Anthropic and leading tech giants have developed the most advanced large language models (LLM). 

In other words, Americans are betting on systems that can give more or less misleading answers to more or less silly questions. Some states in the US have considered introducing AI safety regulations to protect children, but such initiatives are of course opposed by the industry, the Republicans and Trump.

Americans view the situation as a competition between American and Chinese records in AI, as a threat from China, and as a national security issue. The Chinese are not blind to the issue of competition, and of “catching up” with the US, where, for example, DeepSeek’s language model showed that it was not entirely impossible. 

However, the Chinese leadership regards AI as a general-purpose technology, similar to key general-purpose technologies such as the steam engines that transformed economies during the first industrial revolution in the 19th century. According to Xi Jinping’s Marxist terminology, as I have explained elsewhere, AI is an important component of the new quality productive forces that will be able to transform the Chinese economy and society. Perhaps this also includes some Chinese AI models that can provide silly answers to silly questions. But that is not the most important function of AI in China’s future.

Related: The Buddha’s lesson for robots: Who’s real, who’s not? | China’s fluid humanoid stuns — but US still rules robot brains