White House Accuses Moonshot AI of Utilizing Anthropic Technology

Allegations Against Moonshot AI

The White House has accused Moonshot AI of secretly utilizing technology from Anthropic to create its chatbot model, Kimi K3. This allegation arises shortly after Kimi K3 secured the top position in the Frontend Code Arena, sparking ongoing debates about U.S. regulations affecting artificial intelligence.

Details of the Allegations

Michael Kratsios, the director of the White House Office of Science and Technology Policy, stated that the U.S. government has obtained information linking K3’s development to Anthropic’s Fable model. In a post on X (formerly Twitter), Kratsios claimed, “We have information that Moonshot AI distilled Anthropic’s Fable for the development of its K3 model.” However, Kratsios did not provide any technical documentation or evidence to substantiate his claims.

According to Kratsios, Moonshot AI has developed a sophisticated internal platform designed to conduct large-scale distillation against American models. This technology purportedly allows them to adapt their methods quickly, making it difficult for competitors in the U.S. to notice their activities. He elaborated, “Model distillation typically involves training a smaller or less expensive system with responses produced by a more powerful model.” While he acknowledged that developers often use this technique for legitimate purposes, he accused Moonshot AI of engaging in unscrupulous practices.

Kimi K3’s Rise and Its Implications

Kimi K3 recently achieved first place on the Frontend Code Arena, with a benchmark score of 1,679, outpacing Anthropic’s Claude Fable 5, which scored 1,631. Despite this achievement, Moonshot AI has not released detailed information about K3, as its full weights are expected to be disclosed on July 27.

Following the model’s benchmark victory, David Sacks, former White House czar for crypto and AI, expressed concerns that stringent U.S. regulations could inadvertently allow Chinese laboratories to close the performance gap in AI technologies. He emphasized the importance of permitting innovation without excessive governmental oversight.

Broader Context of AI Competition

Bessent, during a Fox Business interview, noted that the United States might impose sanctions on AI companies abroad that are found to engage in theft of intellectual property. Additionally, the New York Post reported that firms such as Anthropic, OpenAI, and Google are worried about unauthorized model distillation from Chinese labs.

As competition intensifies in the AI sector, Chinese firms like DeepSeek continue seeking substantial funding to challenge their U.S. counterparts. Reports suggest that DeepSeek might be valued at approximately $71 billion in prospective funding negotiations.

A Call for Fair Competition

Kratsios underscored the distinction between regular market competition and the actions allegedly taken by Moonshot, stating, “Large-scale, covert industrial distillation aimed at stealing proprietary U.S. technology and undermining American research is unacceptable.” These allegations and their implications are poised to further complicate the evolving landscape of AI competition and regulatory frameworks.