Anthropic accused the Chinese companies Moonshot and DeepSeek of copying the capabilities of their Claude artificial intelligence model. According to the company, the services involved were designated as "transfer stations" and frequently used stolen or fraudulently obtained credentials to access Anthropic's systems.
These transfer stations routed requests that had originally been made by users of Chinese platforms, functioning as intermediaries between users in China and Anthropic's systems in the United States.
The accusation comes in a context of growing tension between Western and Chinese technology companies in the artificial intelligence sector, where the dispute for technology and language model capabilities has intensified in recent years.
Anthropic did not specify exactly what type of information or capabilities were effectively copied, but the company revealed having detected this pattern of unauthorized access through the compromised credentials.
