Anthropic, referred to as the 'rival of OpenAI,' has decided not to provide AI services to Chinese corporations, marking it as the first among American artificial intelligence companies, according to a report by the Financial Times on the 4th.
An Anthropic executive told FT, "This is a measure to prevent Chinese authorities from utilizing advanced AI technology for military and intelligence purposes." He noted that the same principles would apply to U.S. adversaries such as Russia, Iran, and North Korea.
FT reported that this policy will be implemented immediately and is likely to also apply to Chinese tech corporations such as TikTok's parent company ByteDance, Tencent, and Alibaba.
The company stated that this decision will inflict a "hit of hundreds of millions of dollars" on Anthropic's revenue, while emphasizing that "the advancement of AI should align with the interests of free democratic countries and contribute to U.S. AI leadership."
This action comes amid increasing concerns within the U.S. that China is using AI for military purposes, ranging from hypersonic weapons to nuclear weapons modeling.
The previous Biden administration implemented comprehensive export controls to prevent China from securing American AI, but the Donald Trump administration did not announce separate export control measures while pursuing a U.S.-China summit, FT reported.
According to a source familiar with the matter, Anthropic's new policy partially targets Chinese corporations in Singapore, and he said, "This could affect companies like ByteDance, Alibaba, and Tencent."
Founded in 2021 by former OpenAI employees, Anthropic emerged as a rival to OpenAI with the launch of its AI model "Claude," which is among the most widely used AIs globally alongside ChatGPT.
Anthropic recently raised $13 billion (approximately 18.1 trillion won) in funding, increasing its valuation to $183 billion (approximately 255 trillion won). It is reported that Google, which invested from the beginning, holds a 14% equity stake.
In China, the use of American AIs such as ChatGPT, Claude, and Google Gemini is banned. However, FT reported that they can be accessed indirectly through virtual private networks (VPNs).