Anseropic accuses Chinese AI companies of illegally collecting data for its Claude chatbot, with tensions continuing to escalate. Anseropic specifically names Deep Seek, Moon Shadow, and MiniMax, alleging that these companies created大量虚假账户 and collected millions of dialogue records to train their own AI platforms.
It is reported that Deep Seek had 150,000 interactions with Claude, Moon Shadow over 3.4 million, and MiniMax reached 13 million interactions. Although this technique, known as “distillation,” is a common development method, Anseropic’s terms of service explicitly prohibit such actions, including measures to prevent use within China.
At the same time, Anseropic’s statement comes after its competitor OpenAI previously accused Chinese companies for similar reasons. OpenAI has submitted this issue to the U.S. House of Representatives China Committee, criticizing Chinese AI companies like Deep Seek for using “new distillation techniques” to unauthorizedly leverage U.S. technology.
Anseropic warns that such actions could threaten national security or lead to AI being used to develop more advanced military weapons or tools for mass surveillance of American citizens. The company calls for U.S. government intervention and emphasizes that “these activities are becoming increasingly complex, and the window to respond is narrowing.”
However, some critics point out that Anseropic itself has long collected data from the open internet, condemning this move as “hypocritical.” The company responds that distillation involving certified open-source technologies may be legitimate, but foreign laboratories illegally distilling U.S. models could bypass security protections.
Currently valued at $380 billion (approximately 547.2 trillion Korean won), Anseropic is facing multiple lawsuits over unauthorized use of copyrighted data. Last year, it agreed to pay $1.5 billion (about 2.16 trillion Korean won) in a settlement, setting a record for the highest settlement in copyright litigation.