White House Accuses Moonshot AI of Distilling Anthropic Model
AI-generated image · US National Wire
The White House alleges the Chinese lab used Anthropic's Fable 5 to develop its Kimi K3 model.
White House director Michael Kratsios has accused the Chinese AI lab Moonshot AI of illegally distilling Anthropic's Fable 5 model to create its own latest release, Kimi K3, according to reporting from Wired.
Moonshot AI released Kimi K3 last Friday. The model has drawn global attention for capabilities that Wired reports are comparable to leading frontier models from Anthropic and OpenAI. Unlike the proprietary systems produced by U.S. AI giants, Kimi K3 is an open-weight AI system, allowing users to tinker with the model.
The accusation comes amid internal divisions within the Trump administration regarding Chinese AI. Wired reports that the Commerce Department, led by Howard Lutnick, has suggested the situation is less alarming, while other officials are pushing for an executive order to prevent the theft of American AI. While the Commerce Department has utilized export controls to combat these issues, some argue the current situation proves those controls are failing, though others suggest the theft of proprietary information is a separate issue from export failures.