China's Security Chief Warns Against Foreign AI Models After Amodei's Call to Slow Down
Dario Amodei's call to slow the pace of AI drew rare support from both Musk and Altman — but China's state-backed Global Times reads the essay as a move to shut China out of global AI governance, and the country's own security minister…
China's Security Chief Warns Against Foreign AI Models After Amodei's Call to Slow Down
Dario Amodei's call to slow the pace of AI drew rare support from both Musk and Altman — but China's state-backed Global Times reads the essay as a move to shut China out of global AI governance, and the country's own security minister frames American models as a threat to the party. The statements come in the days before expected US–China AI talks and reveal a tension within Beijing's own line.
Anthropic chief Dario Amodei called on Saturday for slowing the development of advanced AI, in the essay "We Must Pace the Frontier". According to a Reuters Factbox, he won rare support from both Elon Musk and OpenAI chief Sam Altman — and triggered a broader debate over so-called AI "doomerism".
China's response: Barriers and monopoly
The state-backed newspaper Global Times responded with an editorial. As relayed by Reuters, the paper claimed that Amodei's real agenda was "attempting to hinder China's AI development through technological barriers and regulatory monopolies, maintain Washington's monopolistic hegemony in advanced technology, and exclude China from the global AI governance system".
One important caveat: Global Times is a state-backed opinion newspaper, and the editorial should not be read as confirmed Chinese government policy. But the tone is clear — Amodei's warning is presented not as a safety question, but as a geopolitical move aimed at China.
The security minister's warning
The next day, the country's security chief and intelligence head, Chen Yixin, went further. In an article in the state-run magazine China Cyberspace on Sunday 14 September, as reported by The New York Times, Chen warned that advanced American models — including Anthropic Mythos and OpenAI GPT-5.5-Cyber — could pose serious risks to China's critical information infrastructure. According to Reuters, he called for a comprehensive strengthening of AI security; according to the Times, he demands more party control over AI and stricter state oversight.
The Times itself points to the tension: Beijing urges the US not to exaggerate the dangers of AI, while the country's own intelligence chief frames the technology as a threat to the Communist Party's security. It is a paradox worth noting, not something that can be resolved here: the diplomatic line discourages doomerism in Washington, while the security apparatus itself uses nearly the same language about foreign models.
Worth noting here as well: neither Chen Yixin's article nor the Global Times editorial is available in the original — all accounts rest on secondary sources, the Reuters Factbox and the Times' coverage.
Trump's response
President Donald Trump pushed back against the very idea of civilian guardrails for AI. "The only control or 'guardrails' that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT, and the U.S.A. has that, in spades!", Trump wrote on Truth Social, according to the Reuters Factbox — that is, the only control AI needs is a strong and smart president, which the US has "in spades". He added that China would benefit from doubt about AI development.
Why now
The reactions come in the days before expected US–China AI talks, ahead of a planned Trump–Xi summit on 24 September, as the Times notes. The statements — Amodei's pace appeal, the Global Times editorial, Chen's warning and Trump's dismissal — thus form the diplomatic backdrop the talks will have to contend with. And the Global Times editorial remains an opinion piece from state-backed press, not a confirmed statement from Chinese authorities.