The strongly worded statement was published on Yuyuantantian, which is said to be a social channel linked to state broadcaster CCTV that often reflects official Beijing policy. China has reportedly set strict terms ahead of an upcoming bilateral summit, declaring that productive artificial intelligence (AI) discussions with the US cannot happen until Washington proves its domestic developers follow equal safety, auditing and disclosure standards. Moreover, Chinese commentators called out US AI developer Anthropic for bad behaviour while laying down ground rules for any future diplomatic meetings.
The commentary stressed that both nations must jointly establish the boundary separating genuine national security hazards from regular market competition, according to a report by Bloomberg. Rising friction between the world’s two largest economies was evident throughout the state media analysis, which singled out Anthropic’s flagship system, Claude, according to the report. In a piece titled “Anthropic Has Contracted the American Disease,” the commentary argued that cutting-edge American models have taken a distorted direction, accusing the US government of repeatedly backing away from its duty to enforce real security boundaries.
The post accused Anthropic’s AI models of overstepping user privacy thresholds, carrying out secret monitoring and transferring web domains without permission.