appConfigEntityId

bodyType

customContents

entityId

entityUuid

isVideo

articlePageQuery

article

articleSchemaContent

content

hooksArticlePageAdvertisementConfigArticle

hooksArticlePageFloatingButtonConfigArticle

appConfig

filter

json

liveArticle

Xinmei Shen

The global AI community has pushed back against the Trump administration’s claims that Moonshot AI’s groundbreaking Kimi K3 distilled US models, citing a lack of justification and arguing that the practice does not constitute intellectual property theft.
In separate posts on X on Thursday, a number of AI experts called the accusations “political” and “reckless”, pointing out that an AI model’s output was not copyrighted.
A Moonshot employee on Thursday hit back at the US officials’ claims that the Chinese company conducted “large-scale, covert industrial distillation” against US models including Anthropic’s cutting-edge Fable 5, by pointing out the narrow window between the release of Kimi K3 and Fable 5.
“Yes, Fable went public on July 1 and K3 launched on July 15. We trained a brand new frontier model in JUST 15 DAYS. Guinness World Record stuff,” Moonshot AI employee Randy Xian wrote on X.
While US officials and AI labs have long attributed Chinese AI’s rapid advancement to distillation, a common industry practice of using the outputs of larger, more capable models to train smaller ones, the White House adviser on Wednesday escalated the accusations by describing the practice as “stealing proprietary US technology”.
Michael Kratsios, US President Donald Trump’s science and technology adviser, made the accusation on X on Wednesday. Then US Under Secretary of State Jacob Helberg echoed Kratsios’ remarks, characterising Moonshot AI’s alleged distillation efforts as “a heist of invaluable American Intellectual Property”.
Anthropic announced Claude Fable 5 in early June, with built-in “safeguards” that downgraded a range of requests, including distillation efforts, to the older model Claude Opus 4.8. The company then took Fable 5 offline following Washington’s directive to block all foreign nationals, before re-releasing it on July 1.
Beijing-based Moonshot did not publicly respond to the latest accusations from Washington and did not immediately respond to a request for comment on Thursday.

Researchers also pointed to the lack of evidence for any distillation claims, while some highlighted Kimi K3’s technical innovations.
It is generally difficult to prove whether a large language model is distilled from another model, and there is currently insufficient evidence to prove if Kimi K3 was distilled from Fable 5, according to Michael Chau, professor in Innovation and Information Management at the University of Hong Kong Business School.
In a Wednesday post on X, Stephen Hsu, professor of Computational Mathematics, Science, and Engineering at Michigan State University, called it “lazy and simplistic” to attribute K3’s advances to distillation of US models.
The Kimi model’s breakthrough “lies in fundamentally re-architecting the model’s computational paradigm through innovations”, according to Hsu, citing the Kimi K3 paper, which mentioned mechanisms such as Kimi Delta Attention and Attention Residuals, “architectural updates designed to improve how information flows across sequence length and model depth”.
Huang Zhenxin, Moonshot AI’s head of enterprise business, was quoted as saying earlier this week by Chinese media outlets that Kimi K3 was “not a distillation or replica of existing models”, and that its performance leap stemmed from “underlying original architectural innovation”. Huang made the comments to local media before Washington’s accusations.
Others criticised the Trump administration’s attempt to categorise distillation as intellectual property theft.
An AI model’s output or chain of thought is not copyrighted nor treated as trade secrets, Kevin Xu, founder of Interconnected Capital.
“Thus, alleging distillation as ‘IP theft’ is not a serious legal argument. It is a political and lobbying talking point masquerading as a legal argument,” Xu said.
US Treasury Secretary Scott Bessent on Thursday threatened action against Chinese AI firms, saying on X that “sanctions and Entity List designations will be on the table” when distillation attacks “cross the line into intellectual property theft”.
Bessent’s threat “appears to be a bit reckless”, as the justification for using sanctions in this case remained “quite unclear”, Paul Triolo, partner and technology policy lead at DGA-Albright Stonebridge Group.
As more US businesses look to open-weight alternatives amid soaring AI spending, Silicon Valley start-ups called for the Trump administration not to restrict access to Chinese open-weight AI.
“Denying American start-ups access to models available abroad would stifle competition, entrench incumbents, and function as a tax on intelligence, raising their costs and narrowing their choices while foreign competitors retain access to the full global market,” 179 US firms, including prominent start-up incubator Y Combinator, said in a letter addressed to Kratsios and Commerce Secretary Howard Lutnick, and obtained by Politico.
Instead of penalising Moonshot’s Kimi models, restricting Kimi access in the US penalises “all American consumers and enterprises”, Xiaoyin Qu, founder of US AI agent firm Tycoon AI, wrote on social media on Thursday.
Such a move would increase American companies’ AI operational spending and slow down AI adoption in the US, she said.


Tesla’s Elon Musk hails Xi Jinping’s leadership and Chinese AI strength

Korean dating show featuring AI and human contestants sparks fascination in China

Hong Kong needs social science to help deliver on its 5-year plan

AI trade outlook brightens on falling oil and Meta’s Muse agent ahead of Xi-Trump meeting

Artificial intelligence (AI)

Global AI experts push back on US ‘distillation’ claims against Moonshot’s Kimi K3 model

A Kimi logo is pictured at the Moonshot AI booth at the World Artificial Intelligence Conference (WAIC) in Shanghai on July 18, 2026. Photo: AFP

The Anthropic logo, a keyboard and a robotic hand are seen in this illustration taken June 5, 2026. Photo: Reuters

Why China’s open-weight AI model Kimi K3 is sparking anxiety in Silicon Valley

People visit the Moonshot AI stand, featuring the Kimi K3 model, during the World Artificial Intelligence Conference in Shanghai on July 18. Photo: AFP

Alibaba says newest Qwen AI model is second only to Anthropic’s Claude Fable 5

With 2.4 trillion parameters, Qwen3.8 is “one of the most powerful models available today”, according to Alibaba. Photo: Shutterstock

The Qwen logo is seen at WAIC in Shanghai on July 17. Photo: AFP

Tech

Tech War

News

China Future Tech

Researchers point to the lack of evidence for any distillation claims, while some highlight Kimi K3’s technical innovations


Researchers point to the lack of evidence for any distillation claims, while some highlight Kimi K3’s technical innovations.


Global AI experts push back on US ‘distillation’ claims against Moonshot’s Kimi K3 model

.css-1c6uqr6{color:inherit;font-weight:inherit;font-size:inherit;font-family:inherit;line-height:inherit;overflow-wrap:break-word;}Researchers point to the lack of evidence for any distillation claims, while some highlight Kimi K3’s technical innovations

Researchers point to the lack of evidence for any distillation claims, while some highlight Kimi K3’s technical innovations