White House backs open AI but draws a hard line at alleged Chinese theft

White House backs open AI but draws a hard line at alleged Chinese theft
The Trump administration is trying to make a difficult argument look simple: it says open AI development is legitimate, but claims Moonshot AI crossed into outright theft while building its fast-rising Kimi K3 model. The result is a new fault line in the AI race, where the fight is no longer just about who has the best model, but how it was trained.

Washington’s position has sharpened quickly. White House tech adviser Michael Kratsios said the administration has information that Moonshot “distilled Anthropic’s Fable for the development of its K3 model” and alleged the company built “a sophisticated internal platform” to evade detection while conducting “large scale distillation against U.S. models.” He added that “large-scale, covert industrial distillation aimed at stealing proprietary U.S. technology and undermining American research is unacceptable.” Treasury Secretary Scott Bessent has taken the same line, warning that “Open source is not open season on American IP,” and that sanctions or Entity List designations could follow if covert distillation amounts to theft.

The administration is also tying the case to export controls. Kratsios has alleged Moonshot accessed Nvidia GB300 systems in Thailand, processors that are restricted for China, raising a second question alongside the IP dispute: not just whether Kimi copied U.S. models, but whether restricted computing infrastructure was used to do it.

But that account is far from settled. Experts cited by TechCrunch say Kimi K3’s strength is hard to explain as a quick copy of Anthropic’s Fable alone. “I don’t think you get a model this strong and this quickly on the heels of Fable doing strictly distillation,” researcher Braden Hancock said, noting the timeline may simply be too short. Others in tech argue the bigger lesson is competitive pressure, not just security risk: “Highly performant open weights frontier models such as Kimi are a competitive threat to OpenAI & Anthropic, but probably for everyone else these are a win.” Open-model advocates have also pushed back on broader crackdowns, with Hugging Face CEO Clément Delangue arguing that “Open models aren’t the risk, they’re the defense!”

That leaves the core divide intact. U.S. officials are drawing a distinction between normal distillation and industrial-scale theft, while critics warn Washington could use that line to squeeze a powerful new open-model rival just as Kimi K3 is reshaping the market.

Continue reading https://foxvector.com/stories/019f9438-8c07-251c-70f7-224719b2d661