China grows anxious over Anthropic’s Mythos ahead of Trump-Xi summit

Chinese officials are growing increasingly anxious that Anthropic's advanced AI model Mythos could be deployed as a weapon against the world's second-largest economy, injecting a volatile new dimension into already-strained U.S.-China relations ahead of a planned summit between President Xi Jinping and President Donald Trump.

AI as a Flashpoint

According to Bloomberg, officials in Beijing are concerned about the cyber capabilities of Mythos and other frontier AI models developed by American companies. They see the potential for such models to be used as offensive tools, and are questioning why Anthropic continues to deny China access to the systems for legitimate commercial purposes.japantimes+1

The concerns are not new. In June, the White House imposed export controls on Mythos after reports that a China-linked group may have accessed the model, prompting Anthropic to pull it from the market entirely. The New York Times reported in May that a Chinese think tank had approached Anthropic representatives in Singapore, urging them to reconsider restrictions on Beijing's access — a request the company declined.semafor+1

Preserving Warmth Before September

China wants to maintain a relatively cooperative atmosphere ahead of Xi's planned September 24 visit to the United States, according to people familiar with the matter. Beijing is seeking a "win-win" approach at AI-focused talks expected to precede the leaders' meeting.japantimes

The upcoming summit would follow Trump's three-day state visit to Beijing in May, when AI and cybersecurity emerged as central topics of discussion. The Carnegie Endowment for International Peace noted that the two sides are broadly "angling for three years of stability," with trade, technology, and Taiwan as the defining issues.scmp+2

Deepening Suspicion

Despite diplomatic overtures, suspicion runs deep on both sides. Washington remains concerned that Chinese companies are using American frontier models to train their own versions through a process known as distillation. Anthropic CEO Dario Amodei has publicly warned of what he calls China's "open-source threat" and "Mythos-class" cyber risks.reddit+2

China, for its part, views the latest generation of American AI models as a considerable leap in technological capability that could represent a threat to its national interests. With the September summit approaching, Mythos has become a symbol of the broader contest over who controls the most powerful AI systems — and what limits should govern their use.nytimes