China accuses CEO of one of America's biggest AI companies of bad behaviour

Chinese state media has accused US AI company Anthropic of unethical practices ahead of upcoming bilateral AI talks between the two nations. State broadcaster CCTV's affiliated social media channel Yuyuantantian argued that Washington must verify its domestic AI companies align with safety, disclosure, and auditing standards before any meaningful cross-border discussions can take place.

The commentary emphasized the need to distinguish between genuine security threats and mere technological competition, stating that all parties must agree on this boundary.

Anthropic's flagship AI model, Claude, was singled out for allegedly violating user privacy, covertly monitoring users, and transmitting external web domain records without permission. The report claimed that leading US AI models have strayed from a proper path due to a lack of enforcement of baseline safety measures by Washington.

Beijing accused the US of using its domestic safety rules to dictate global digital standards, framing American regulatory policies as a means to hinder Chinese technological advancement and force acceptance of rules set by private US tech corporations.

The dispute highlights growing concerns among Chinese officials about advanced Western AI models, particularly Anthropic's Mythos system, which they fear could be used offensively against national infrastructure. This apprehension stems from the fact that Anthropic has blocked Chinese-controlled organizations from accessing its platforms, citing US national security priorities.