Artificial intelligence (AI) technology has the potential to revolutionize the way the world works and communicates. AI also plays an outsized role in technological competition between the United States and China, in what some call the “AI race.” What is China’s current stage in the AI race as compared to the United States, and what challenges and risks lie ahead in adopting AI technology?
Hi, my name is Jeff Ding. I’m an assistant professor of political science at George Washington University. On the side, I publish a newsletter called China AI, which features translations of Chinese language writing on AI related topics. I’m going to start by talking about public perceptions of the U.S.-China rivalry in AI.
What is the public perception toward AI rivalry between the U.S. and China?
I think there are two main points to highlight here. One of the main public perceptions is that AI has become a central front of great power competition between the US and China. And I do think there’s a lot of historical precedent for this. Historically, the rise and fall of great powers has generally proceeded from a period of significant technological change, which then creates an opportunity for one great power to sustain economic growth at higher levels than its rivals. And then eventually, that great power converts a sustained growth advantage into geopolitical and military influence. And we’ve seen that play out throughout the centuries, with the first industrial revolution, and inventions like the steam engine, and innovations in cotton textiles that propelled Britain to global leadership, followed by the US as rise in the late 19th century, off the back of innovations in electricity, chemicals and machine tools. And even in the competition between the US and Japan in the 1980s, when many predicted that Japan’s dominance of new innovations in semiconductors and consumer electronics would lead to Japan becoming the number one power on the global stage. So I think that public perception is right on. Another key perception and key assumption about US China rivalry and AI is that China is either ahead or very close to overtaking the US in terms of leadership in this cutting-edge technological domain.
In the next section, I’ll talk about why I think that public perception is flawed, and the reasons why it is flawed.
What is the Chinese perspective on the danger of AI?
Regarding this question of China’s perspectives on the danger of AI, what I want to underline is that Chinese actors definitely grapple with risks and fears about AI development similar to a lot of different institutions in the West. And I think this is an area for continued collaboration. One thing that I oftentimes have to remind myself, as we’re bombarded by news of just an increasingly conflictual relationship is that cooperation is happening all the time, and it is happening in the space.
For example, there is a joint technical committee of two of the most important international standard setting organizations that is working on setting technical standards for AI development, that have some bearing on controlling and checking against the dangers of AI. And Chinese stakeholders, Chinese company participants, Chinese government participants are involved are very much involved in the standards setting efforts. For example, in that joint technical committee, Chinese stakeholders and representatives lead a working group on the controllability of automated systems. So that’s an area where you have people on the ground working together to cooperate on these issues.
I think one very interesting aspect of all of this is US decision makers and US thinkers have pointed to the need to find a permissive action link for AI. And what that permissive action link refers to is a technology that guards against unauthorized use of nuclear weapons. It was an electronic lock system that would ensure that a nuclear bomb cannot be detonated without express approvals. And the US even admits the fiercest geopolitical rivalry with the Soviet Union during the heart of the Cold War, the US still found ways to try to share information and cooperate with the Soviet Union on this permissive action link technology.
And so, I think it’s a positive sign that US decision makers are looking for areas where it wouldn’t make sense for both sides even as they compete as rivals where it would make sense for them to cooperate to control the dangers of powerful technologies, whether that be nuclear weapons, or transformative AI systems.
What’s China stage in the AI race right now?
So, in terms of China’s current stage, in this competition over AI, and remember from the first question, I emphasize how the core stage are a core step in terms of great power competition in these emerging technologies is how they affect each country’s ability to sustain economic growth. And so, where China sits in terms of being able to leverage AI, to improve productivity across the entire economy, because that’s what’s so incredible about these general-purpose technologies like AI, and electricity. Where China sits, I think, is very far behind the US. And I think the US has a clear leadership position in terms of being able to capitalize on AI to sustain productivity growth.
I think we see that in terms of cutting-edge innovations, like new large language models, such as opening eyes chat GPT, where there is still a significant gap between chat GPT and models like Baidu’s ERNIE bot that are trying to catch up in the space. And I think this links back to this historical tendency of the US to over inflate and overstate their technological rivals’ capabilities in new technologies. I think another factor that points to sustain us leadership is the US has dominance of some of these foundational building blocks of AI developments, such as the hardware that is used to train and run algorithms, and also the foundational open-source software, much of which is developed in the United States.
One point I want to emphasize is, oftentimes when we compare China and the US in terms of AI capabilities, we overly focus on innovation capacity. And we neglect this idea of diffusion capacity, which I highlighted in a recent article for the review of international political economy. And diffusion capacity refers to a country’s ability to spread and embed new technological advances throughout a wide range of productive processes in its economy. And in recent testimony to the US China economic and security review commission, I highlighted why China still is far behind the US in terms of its diffusion capacity, we see that with China and China lagging behind other countries in terms of adopting information and communication technologies like cloud computing, and even computers for business usage, I think we oftentimes ignore the fact that beyond the most developed clusters like Beijing, once you get into third and fourth tier cities, and even farther below, and beyond the coastal areas, the diffusion of these digital technologies is still very, very slow in China. And in my testimony, I show why, how these slow diffusion trends seem to also apply to AI.
This transcript has been lightly edited for clarity. Please refer to the video interview to ensure accuracy.