Breaking
ARThe Yemeni government announces a military operation against the Houthis and the Houthis announce attacks on Saudi AramcoBRProfessor Dorinha votes in Palmas and awaits election resultsRUIn Podolsk, a motorcyclist died in an accident with a LexusCNLate-night fire in Xinfeng burned 2 buildings, 6 people were trapped and rescuedBRPolice record 15 incidents during voting in Paraíba, including arrests for violating vote secrecyBRSchool prepares to welcome voters on voting dayBRCivil Defense issues warning of rain with lightning and strong winds in Baixada Santista during the 2026 ElectionsUKChrista Pike remains critically ill after failed execution attemptsRUCara Delevingne attended the Tom Ford show in a revealing corduroy suitRUNear Yaroslavl, a woman was covered with bricks during an explosion at a pyrotechnics plantARThe Yemeni government announces a military operation against the Houthis and the Houthis announce attacks on Saudi AramcoBRProfessor Dorinha votes in Palmas and awaits election resultsRUIn Podolsk, a motorcyclist died in an accident with a LexusCNLate-night fire in Xinfeng burned 2 buildings, 6 people were trapped and rescuedBRPolice record 15 incidents during voting in Paraíba, including arrests for violating vote secrecyBRSchool prepares to welcome voters on voting dayBRCivil Defense issues warning of rain with lightning and strong winds in Baixada Santista during the 2026 ElectionsUKChrista Pike remains critically ill after failed execution attemptsRUCara Delevingne attended the Tom Ford show in a revealing corduroy suitRUNear Yaroslavl, a woman was covered with bricks during an explosion at a pyrotechnics plant
Back美財長貝森特:中國AI模型Kimi竊取技術反洩露解放軍機密
美財長貝森特:中國AI模型Kimi竊取技術反洩露解放軍機密
Tech
自由时报2 hours agoTech2 min readChinaView translation

美財長貝森特:中國AI模型Kimi竊取技術反洩露解放軍機密

美國財長貝森特指出,中國AI模型「月之暗面」Kimi K3在竊取美國模型技術時,意外將中國解放軍武器計畫傳輸給美國Anthropic公司。

Quick Look

美國財長貝森特表示,中國AI模型「月之暗面」Kimi K3透過工業化蒸餾竊取美國技術,卻意外自稱是Claude,並將中國解放軍的武器計畫與監控數據傳輸給美國開發商Anthropic,造成機密外洩。

AI-generated summary

Why It Matters

Anthropic於9月10日發布報告,指出中國AI公司在蒸餾外國模型過程中意外導致機密外洩。貝森特在Axios節目中引用此事件討論美中AI競爭。

Font size

美國政治新聞網站《Axios》報導,美國財長貝森特受訪時透露,中國的AI模型竊取了美國AI模型的技術,結果卻出現了令北京當局尷尬的局面:被視為中國最強大的AI模型─「月之暗面」(MoonShot AI)的Kimi K3,有時竟然自稱是美國AI模型Claude;更出乎意料的是,它還把中國解放軍的武器計畫拱手送給了Claude的開發商,反倒給美國「幫了個大忙」。

貝森特在10月3日播出的《Axios Show》節目中談及美中的AI競爭,表示在AI方面,自己最擔心的事情之一就是美國被中國反超。在談到AI模型的開發時,貝森特說,中國的開源模型會進行「工業化蒸餾」(industrial distillation),這個術語其實就是「竊取美國模型」的一個好聽說法。

貝森特隨後點名「月之暗面」的Kimi K3模型,並提到了該模型的1個最大缺陷。他說,「Kimi K3認為自己是美國Anthropic開發的Claude,有時候Kimi會直接告訴你它是Claude。」

貝森特還爆料,Kimi K3向美國洩露了中國的軍事祕密。他說:「這件事已經公開了,因此我可以談。Kimi k3把一些中國解放軍的武器計畫傳給了Anthropic,這可真是『幫了大忙』。」

Anthropic於9月10日發布報告指出,在中國AI公司蒸餾外國AI的過程中,意外曝光了中國或中企內部機密。例如,1名中國軍方用戶使用Kimi K3模型對監控數據進行分析,他從閉路電視監控錄影庫中加載了針對特定目標人物的監控數據,要求Kimi K3分析這些閉路電視數據,以了解被跟蹤者的行為是否異常。

這些閉路電視數據包括來自成都數百個攝影機的監控影像,其中包括中國解放軍設施外的攝影機。結果「月之暗面」透過中介網路,把這些數據傳輸給了Anthropic的Claude模型,要求其作答,因此就將中國軍方的一些機密資料外洩。

Open Questions

  • 中國政府是否會對月之暗面採取懲罰措施?
  • 未來是否會有更多類似的蒸餾模型洩密事件?

Related Topics

This article was originally published by 自由时报.

Related Stories

Trump appoints National Intelligence Chief Clayton to lead the newly established interdepartmental task force "Superintelligence Force" and also serve as artificial intelligence czar
BREAKING·

Trump appoints National Intelligence Chief Clayton to lead the newly established interdepartmental task force "Superintelligence Force" and also serve as artificial intelligence czar

U.S. President Trump appointed Clayton, the current chief of national intelligence, to lead the newly established interdepartmental task force "Super Intelligence Force" and also serve as the artificial intelligence czar. It aims to coordinate the federal government's AI policy and promote AI as a core strategic asset for national competitiveness and geopolitical confrontation. The team will assess the opportunities and risks of AI to national security and the economy within 120 days, and submit an inspection report.

自由时报
2 min read
Political Censorship and “Soul Struggle” of Artificial Intelligence in China: From the Wu Shufang Case to Global Spread
Tech·

Political Censorship and “Soul Struggle” of Artificial Intelligence in China: From the Wu Shufang Case to Global Spread

Separate tests by Taiwanese media and the Christian Science Monitor found that Chinese AI models censored, refused to answer, or output biased content when dealing with the history of the Cultural Revolution and political issues. Research points out that these models with censorship mechanisms are being widely adopted by global enterprises through API interfaces.

德国之声
4 min read
Indian media: China is emerging as a technological power with its manufacturing capabilities and state support
Tech·

Indian media: China is emerging as a technological power with its manufacturing capabilities and state support

An article in India's Deccan Herald pointed out that China, with its strong manufacturing capabilities, national policy support and forward-looking supply chain layout, occupies a global dominant position in the fields of electric vehicles, batteries, industrial robots and renewable energy, and is steadily moving towards becoming a technological power.

中国新闻网
1 min read
China's first "four-car simultaneous movement" cross-railway rack dismantling process was successfully implemented at Chongqing Science City Station
Tech·

China's first "four-car simultaneous movement" cross-railway rack dismantling process was successfully implemented at Chongqing Science City Station

China Railway Construction Group successfully implemented the country's first "four-vehicle simultaneous movement" cross-rail belay frame dismantling process at the Chongqing Science City Station of the Chengdu-Chongqing Middle Line High-speed Railway. This innovative construction method significantly shortens the construction period, marking the project entering the final stage of decoration and renovation, and is expected to reach full debut conditions by the end of October.

中国新闻网
2 min read
South Korean financial institutions were attacked by cyberattacks, and authorities did not rule out the possibility of AI involvement
Urgent·

South Korean financial institutions were attacked by cyberattacks, and authorities did not rule out the possibility of AI involvement

South Korea has recently experienced a number of personal data leaks and cyber attacks at financial institutions. Lee Oh-won, chairman of the Financial Services Commission, warned that the attacks may involve artificial intelligence and has ordered a comprehensive upgrade of the security structure and launched an investigation. The source of attack traffic spans many countries, and authorities are expanding their investigation of vulnerabilities in the financial industry.

中央社
2 min read
More on this topic月之暗面