Anthropic has decided to stop offering artificial intelligence services to companies whose majority of shares are owned by Chinese organizations. This is considered the first such decision made by a U.S.-based AI company.
According to one of the executives of the San Francisco-based company, who spoke to the Financial Times, the primary goal of this decision is to prevent Beijing from using the technology for military and intelligence purposes. This policy could potentially include major companies such as ByteDance, Tencent, and Alibaba.
The unnamed executive stated that this move “closes a gap that allows Chinese companies access to the most advanced AI.” He also mentioned that the policy will extend to other U.S. rivals, including Russia, Iran, and North Korea.
Anthropic emphasized that this step is intended to “ensure the development of flexible AI capabilities aligns with U.S. leadership and democratic interests.”
The company acknowledges that this decision could result in hundreds of millions of dollars in lost global revenue. Nevertheless, they stress that, despite the risk of losing customers to competitors, they want to highlight the seriousness of the issue.
Concerns in the U.S. are growing over Chinese companies attempting to gain access to American technology with less oversight by setting up subsidiaries in countries like Singapore. The obligation of Chinese companies to provide data to the state at any time is viewed as a national security risk for the United States.
This move comes amid worries that China could use AI in various military domains, from hypersonic weapons to nuclear modeling.
Meanwhile, China’s startup DeepSeek released its open-source R1 model earlier this year, which was considered competitive with American models. OpenAI has claimed that DeepSeek illegally accessed its models to train the R1 system.
