US-China AI War: Are Rapid Advancements in AI Dangerous?

Over the recent few years, AI has become easier to access. And I could feel how quickly AI has developed. New models are constantly being released, and I think functions that were difficult for AI to perform are becoming possible in short amounts of time. But these days when I look up news on AI, there are more articles about the competition between the U.S. and China rather than about competition among AI companies. What I really wanted to know while reading various articles was not which country was “winning” the AI race but what happens to AI safety when the competition moves too quickly. Even when some proposed to slow down AI development, I wondered if it is realistically possible for one country to slow down if the other side doesn’t.

The Guardian explains how Chinese AI companies compete with those in America. Kimi K3 released by the Chinese startup Moonshot is free to use and because it’s an open-weight model, users can download it directly. The article states that Chinese companies are competing by putting forth such cheap and public models. At first, this seemed a bit unrelated to safety concerns. However, if more people are more easily using powerful AI, then price and accessibility cannot be viewed separately from AI safety issues.

The article I found on Al Jazeera fascinated me more. Dario Amodei, CEO of Anthropic, states that we need to slow down the development of powerful AI to reserve the time to manage potential threats. At the same time, he also expressed the need to limit China’s access to advanced AI technology and chips. On the other hand, the Chinese government criticized that such an approach is America’s excuse to maintain its status in the AI race. This seemed the most complicated for me. If AI is indeed developing so quickly that it’s dangerous, then there is the need to adjust the speed of development. But if there is competition between two nations, then neither side would want to be the first to slow down.

Figure 1. An NVIDIA H100 high-performance GPU used for AI computing. Image by Geekerwan.

Chinese experts on Global Times claimed that if the U.S. focuses too much on competing with China in AI, then discussions on safety and long-term risks could be pushed back. They also expressed that if the competition is viewed in terms of winning and losing, it may become more difficult for countries to create common rules and cooperate. But because Global Times is a news outlet strongly reflecting China’s perspective, I believe that it’s more appropriate to treat it as an article that shows how China views America’s AI policies rather than as an objective source.

“Problems arise when speed becomes the main focus, and the time to check for safety and create common rules is ignored.”

I also do not think the U.S. and China should stop competing, because such competition encourages technological advancement and helps with creating cheaper services. Problems arise when speed becomes the main focus, and the time to check for safety and create common rules is ignored. Risks like AI misuse or cyberattacks happen internationally, beyond borders. Even if a country wins the competition, the risks don’t disappear. Therefore, I think that the two countries can compete in the AI market, but they need to collaborate on issues that impact them both. They need to at least share information and common safety protocols for concerns like major AI accidents and abuse.

After reading the three articles, what concerned me most was not the speed of AI development but the fact that both the U.S. and China are in situations that make it difficult for them to slow down first. If both countries continue to think that if one side slows down, then the other will win, it will be difficult for AI safety to be prioritized. Therefore, I think that the most important thing amidst the AI War will be less about who is faster but more about how much the two will work for the common good and safety while remaining rivals.

Figure 2. U.S. Representative Ro Khanna meets with Chinese Premier Li Qiang in Beijing on September 21, 2025. The delegation discussed international AI safety standards. Photo by the Office of Representative Ro Khanna.

Works Cited

• Al Jazeera Staff and Reuters. “‘Silent Cold War’: Why Calls to Slow AI Have Sparked New US–China Frontier.” Al Jazeera, 14 Sept. 2026.

• Geekerwan. “NVIDIA H100 (极客湾Geekerwan) 001.” Wikimedia Commons, 14 June 2023.

• Montgomery, Blake. “Headaches for Silicon Valley as China Chips Away at the US’s Lead in the AI Race.” The Guardian, 21 July 2026.

• Office of Representative Ro Khanna. “Representative Ro Khanna Meeting Premier of China Li Qiang (2025).” Wikimedia Commons, 21 Sept. 2025.

• “US AI Race Mind-Set, Fearmongering over China Risk Fragmenting Global Governance amid AI Slowdown Debate: Chinese Experts.” Global Times, 14 Sept. 2026.