China warns AI could threaten party control as it closes the gap with the US

China warns AI could threaten party control as it closes the gap with the US

Chinese AI is catching up with the US, according to Stanford HAI, even as Beijing hedges internally: according to a Times of Israel post, security minister Chen Yixin has described AI as a threat to the Communist Party's grip on power — right before Xi Jinping's expected visit to Washington.

China increasingly presents itself as a counterweight to the US in artificial intelligence, and this year appears to be erasing the American lead. That is the conclusion of experts and a report from the Stanford Institute for Human-Centered Artificial Intelligence, as reported by AP News [a5ab5bbc]. At the same time, a blogger writes in the Times of Israel, China's security minister Chen Yixin published an article in mid-September describing AI as a threat to the Communist Party's hold on power [69afe7dd]. The two signals — growing capability and growing unease in the security apparatus — land in a week marked by diplomatic escalation: Xi Jinping is expected in Washington this month to meet Donald Trump, and both countries have voiced rising concern about AI, according to AP [a5ab5bbc].

A hardening risk framework

What separates Beijing's rhetoric from the American one is what the risk is tied to. According to the Times of Israel post, written by Joe Nalven, Beijing is not waiting to act — but the framework is different: while American AI leaders have warned that the technology could escape human control, the security minister frames the risk primarily as a threat to the party itself. At the same time, Beijing says it supports global AI governance while refusing to slow the strategic race — a position the post's author sets against Trump's statements that safety routines are possible [69afe7dd]. The dating of this signal is worth noting: on September 14, a spokesperson for China's Foreign Ministry is said to have dismissed the warnings from American AI chiefs as fear mongering, according to the same post [69afe7dd].

We emphasize that both statements come from a single opinion post, and that the underlying article by Chen Yixin has not been verified by us.

Regulatory drumbeats

The soft signals have concrete counterparts. The Financial Times has reported that Chinese regulators led by the commerce ministry have consulted leading AI companies including Alibaba, ByteDance and Zhipu on two questions: restricting the transfer of training data abroad, and whether foreign users should still be able to freely download model weights from Chinese AI systems. That is reported in an op-ed in The Hindu [688bc8a8]. The same op-ed estimates — as the writer's own assessment, not an official statistic — that 820 large language models were registered with China's cyberspace authority in early 2026 [688bc8a8].

Taken together, this points to a development in which China's openness strategy — freely available model weights have been a competitive advantage — is being weighed against security and control considerations. None of the cited sources confirm that any change has actually been adopted.

What is verified — and what is not

The most important thing to understand about this story is what rests on solid sources and what does not. That China is closing the AI gap with the US rests on AP and the Stanford HAI report [a5ab5bbc]. The consultations on training data and model weights were reported by the Financial Times, via The Hindu [688bc8a8]. But the claims about Chen Yixin's article, the Foreign Ministry's "fear mongering" remarks, and Beijing's position on global governance come solely from an opinion post in the Times of Israel [69afe7dd], with the primary source unavailable. None of the cited sources document named Chinese regulations, technical standards, or research programs for handling AI that "escapes human control."

The big picture

Many of the closer details — the date of Chen Yixin's article, what concrete steps Beijing may adopt, and what the Xi–Trump summit actually produces — remain open questions. What the sources leave behind is a clearer pattern: China is emerging as an AI counterweight with growing capability [a5ab5bbc], while its own security elite, according to secondary sources, is beginning to describe the technology as a matter of internal power [69afe7dd]. It is at the intersection of these — growing capability, growing demand for control, and concrete regulatory consultations [688bc8a8] — that the Chinese AI-risk debate is taking shape, with the Washington visit as the next natural test.