The American companies building artificial intelligence systems are loudly complaining that their Chinese competitors are unfairly copying their technology, and they are pleading with officials to do something about it.

开发人工智能系统的美国公司正在大声抱怨，指责他们的中国竞争对手在不公平地抄袭其技术，并正恳求官员们采取措施予以应对。

On June 10, Anthropic sent a letter to Senators Tim Scott and Elizabeth Warren, accusing the Chinese tech giant Alibaba of surreptitiously copying its A.I. technologies using a technique called distillation.

6月10日，Anthropic致信参议员蒂姆·斯科特和伊丽莎白·沃伦，指责中国科技巨头阿里巴巴利用一种被称为“蒸馏”的技术，秘密抄袭其人工智能技术。

Like other Chinese companies, Alibaba tapped into Anthropic’s technologies through tens of thousands of unauthorized accounts, according to the letter, which was viewed by The New York Times. Then it used the data it collected to train its own A.I. systems. Anthropic asked the lawmakers, who lead a Senate committee that was about to hold a hearing on A.I., to explore ways of curbing China’s distillation.

据《纽约时报》看到的这封信函显示，与其他中国公司一样，阿里巴巴在通过数万个未经授权的账户使用Anthropic的技术。随后，它利用收集到的数据训练自己的AI系统。Anthropic请求这两位的议员共同探讨如何遏制中国的模型蒸馏。两人领导着一个即将就AI举行听证会的参议院委员会。

“These distillation attacks are carried out illicitly, systematically and at industrial scale to harvest U.S. A.I. capabilities across frontier labs and repackage them as their own,” Anthropic told the two senators, referring to companies on the frontier of A.I. development.

“这些蒸馏攻击是以非法、系统化且工业级的规模进行的，旨在从各家前沿实验室中窃取美国的AI能力，并将其重新包装为自己的成果，”Anthropic在给两位参议员的信中写道，这里所指的是处于人工智能开发最前沿的公司。

Experts say China trails the United States in A.I. development by just six months. Anthropic and other U.S. companies argue that without help from distillation, China would be much further behind, which could affect major A.I. uses like business planning, drug research, mass surveillance and military weapons.

专家表示，中国在人工智能开发方面仅落后美国大约六个月。Anthropic和其他美国公司辩称，如果没有蒸馏技术的帮助，中国会落后得更远，这可能会影响到商业规划、药物研究、大规模监控以及军事武器等AI的核心应用领域。

Their complaints have new urgency now that the Chinese start-up Z.ai has released an A.I. model, GLM-5.2, that is nearly as powerful as the top American systems. It rivals them when used for cybersecurity, an area that American A.I. companies and the Trump administration have singled out as vitally important to geopolitics.

随着中国初创公司智谱AI(Z.ai)发布了一款实力几乎与美国顶级系统齐平的AI模型GLM-5.2，他们的抱怨变得愈发迫切。该模型在网络安全应用方面可与美国系统相媲美，而这一领域正是美国AI公司和特朗普政府强调的对地缘政治至关重要的领域。

But what exactly is distillation, and are Chinese companies the only ones doing it? Here is an explanation.

但究竟什么是“蒸馏”？难道只有中国公司在这么做吗？以下是相关解析。

Is distillation a new concept?

蒸馏是一个新概念吗？

Not at all. Distillation has been common in the tech industry for more than a decade. A small team of Google researchers first developed the technique in the early 2010s as a way of building more efficient A.I. systems.

绝非如此。蒸馏技术在科技行业普遍存在已超过十年。谷歌的一个小型研究团队在2010年代初期首次开发了这项技术，以此来构建更高效的AI系统。

Through distillation, researchers can collect data from a particularly powerful system and use that data to build a system that can run on less expensive hardware.

通过蒸馏，研究人员可以从一个特别强大的系统中收集数据，并利用这些数据来构建一个可以在更便宜的硬件上运行的系统。

The first A.I. model essentially shows the second model how to behave, said Geoffrey Hinton, a former Google researcher who helped develop the technique. “Think of one model as the teacher and the other as a student,” he said.

曾参与开发该技术的前谷歌研究员杰弗里·辛顿表示，第一个AI模型本质上是在向第二个模型示范如何表现。他说：“可以把一个模型看作老师，另一个看作学生。”

Distillation is a way to copy your own A.I.?

蒸馏是一种复制自身AI的方式？

Correct. But some companies used distillation to mimic technologies built by other A.I. labs. They often copied the behavior of open source technologies — systems that anyone can use, modify and copy for free and largely without restriction.

没错。但一些公司利用蒸馏来模仿其他AI实验室构建的技术。它们经常复制开源技术的行为——开源技术是指任何人都可以免费使用、修改和复制，且基本上不受限制的系统。

That is what labs hope to encourage when they open source their systems. The idea is that everyone benefits because A.I. is developed more quickly.

这正是各家实验室在开源其系统时所希望推动的。其核心理念是，由于人工智能能够更快速地发展，每个人都能从中受益。

When is distillation a problem?

蒸馏在什么时候会成为一个问题？

Anthropic, OpenAI and other A.I. labs get annoyed when companies use distillation to mimic the behavior of their proprietary systems — technologies that are not open source. These are typically their most powerful systems.

当其他公司利用蒸馏来模仿其专有系统（即非开源技术）的行为时，Anthropic、OpenAI和其他AI实验室就会感到恼火。这些系统通常是它们最强大的系统。

Anthropic and OpenAI do not allow distillation for their leading systems under their terms of service. But distilling these systems is still common.

根据服务条款，Anthropic和OpenAI不允许对其领先系统进行蒸馏。但对这些系统进行蒸馏的现象依然屡见不鲜。

In April, while testifying in a federal trial in Oakland, Calif., Elon Musk acknowledged the practice at his A.I. company, xAI. When a lawyer asked if xAI had ever distilled technology from OpenAI, Mr. Musk replied: “Generally A.I. companies distill other A.I. companies.”

今年4月，埃隆·马斯克在加利福尼亚州奥克兰的一场联邦审判中作证时，承认了他的AI公司xAI存在这种做法。当律师询问xAI是否曾对OpenAI的技术进行过蒸馏时，马斯克回答道：“AI公司一般都会蒸馏其他AI公司的技术。”

埃隆·马斯克在加利福尼亚州奥克兰的联邦法院，他于4月在此承认xAI使用了蒸馏技术。

Is that illegal?

这么做违法吗？

That’s not clear, said Sarah Tishler, a partner at the law firm Beck Reed Riden who specializes in trade-secret litigation.

专门从事商业秘密诉讼的贝克-里德-莱登律师事务所合伙人萨拉·蒂什勒表示，目前这一点还不明确。

Some legal scholars argue that the practice violates the Defend Trade Secrets Act, a 2016 law that allows businesses to sue over the theft of trade secrets, but courts have not explicitly decided that.

一些法律学者辩称，这种做法违反了2016年出台的《保护商业秘密法》，该法律允许企业就商业秘密被窃取提起诉讼，但法院尚未对此做出明确裁决。

Copyright law does not necessarily apply because distillation is an effort to copy the behavior of the system, as opposed to copying text verbatim.

版权法并不一定适用，因为蒸馏是为了复制系统的行为，而不是逐字逐句地复制文本。

Are the Chinese doing something similar?

中国人也在做类似的事情吗？

It is also not completely clear what Chinese companies are doing. They have likely distilled proprietary models in much the same way that American companies like xAI have done.

中国公司具体在做什么也并非完全透明。它们很可能以与美国公司（如xAI）极为相似的方式对专有模型进行了蒸馏。

Chinese distillation efforts, however, have caused far more concern among Anthropic, OpenAI and the other U.S. companies.

然而，中国的蒸馏举动在Anthropic、OpenAI等美国公司中引发了深得多的担忧。

About 18 months ago, the Chinese start-up DeepSeek shocked Silicon Valley when it showed that it could build effective A.I. far more affordably than many of its American counterparts. OpenAI soon accused DeepSeek of distilling its technologies.

大约18个月前，中国初创公司DeepSeek向硅谷展示了它能以比许多美国同行低得多的成本构建出高效的AI，这引发了硅谷的震动。OpenAI很快便指责DeepSeek蒸馏了其技术。

In February, Anthropic accused DeepSeek and two other Chinese start-ups of improperly harvesting large amounts of data from its systems. Anthropic said the start-ups had used about 24,000 accounts to generate over 16 million conversations with its Claude chatbot that could be used to teach skills to their own chatbots.

今年2月，Anthropic指责DeepSeek和另外两家中国初创公司从其系统中不当获取了大量数据。Anthropic表示，这些初创公司利用了大约24000个账户，与其Claude聊天机器人进行了超过1600万次对话，这些对话可用于向它们自己的聊天机器人传授技能。

How does Anthropic know this?

Anthropic是如何得知这一点的？

Anthropic closely monitors how people use its systems. Certain repeated behavior, the company said, showed that accounts linked to China were lifting data from its proprietary models.

Anthropic密切监控着人们对其系统的使用情况。该公司表示，某些重复性的行为表明，与中国有关联的账户正在从其专有模型中窃取数据。

Anthropic claimed that various Chinese companies had used a network of accounts to gain access to its systems. Each Chinese company, Anthropic said, uses this data to help train its own technologies.

Anthropic声称，多家中国公司利用了一个账户网络来接入其系统。Anthropic指出，每家中国公司都利用这些数据来帮助训练它们自己的技术。

Can Anthropic prevent this?

Anthropic能阻止这种行为吗？

Anthropic, OpenAI and Google are sharing information that they can all use to combat the practice, they said. But it can be difficult to stop. If Anthropic shuts down too many accounts, it may end up barring legitimate users.

Anthropic、OpenAI和谷歌表示，它们正在共享信息，以便共同打击这种行为。但这很难杜绝。如果Anthropic封禁了过多的账户，最终可能会把合法的用户也拒之门外。

Even if U.S. law did bar illicit distillation, Ms. Tishler said, it would most likely have little effect on behavior in China.

蒂什勒指出，即使美国法律确实禁止非法蒸馏，也很可能对中国境内的行为产生不了什么影响。

“So much of this conduct is happening outside the United States,” she noted. “It would be very challenging to address it through a U.S. court.”

“这种行为有太多是发生在美国境外了，”她指出。“通过美国法院来解决这一问题将会极其困难。”

(The Times sued OpenAI and Microsoft in 2023, claiming copyright infringement of news content related to A.I. systems. The two companies have denied those claims.)

（《纽约时报》于2023年对OpenAI和微软提起诉讼，指控其在人工智能系统相关的新闻内容上存在版权侵权行为。这两家公司均否认了这些指控。）

Anthropic5月在旧金山展示AI智能体Claude Cowork。

What else can U.S. companies do?

美国公司还能做什么？

Anthropic called on Congress to pass legislation that would allow “deeper collaboration to combat distillation attacks, both between the U.S. government and leading frontier labs as well as between the frontier labs themselves.”

Anthropic呼吁国会通过立法，允许“在美国政府与领先的前沿实验室之间，以及前沿实验室彼此之间开展更深入的合作，以打击蒸馏攻击”。

The company also said the U.S. government should extend its efforts to limit China’s access to the specialized computer chips needed to train A.I. technologies. The world’s most powerful chips are designed by American companies, and the federal government has used export controls to stem the flow of those chips to China. It is difficult to do distillation without those chips.

该公司还表示，美国政府应当加大力度限制中国获取训练AI技术所需的专用电脑芯片。世界上最强大的芯片均由美国公司设计，联邦政府一直在利用出口管制来遏制这些芯片流入中国。没有这些芯片，进行蒸馏也会变得很困难。

Alibaba declined to comment on Anthropic’s letter to the two senators. Ms. Warren, Democrat of Massachusetts, also declined to comment. Mr. Scott, Republican of South Carolina, did not respond to a request for comment.

阿里巴巴拒绝就Anthropic致两位参议员的信函发表评论。马萨诸塞州民主党参议员沃伦同样拒绝置评。南卡罗来纳州共和党参议员斯科特未回应置评请求。

Would a distillation crackdown have an impact?

打击蒸馏会产生影响吗？

Many experts believe that a crackdown on Chinese distillation would have little effect, and that distillation alone cannot build a top A.I. system as Z.ai did.

许多专家认为，打击中国的蒸馏行为不会有什么效果，而且单凭蒸馏技术是无法构建出像智谱AI那样顶尖的AI系统的。

Others believe that distillation will become less important as companies build systems, like GLM-5.2, that are designed to serve as A.I. agents. Training these agents — digital assistants that can use other software to perform tasks — is much harder to duplicate through distillation.

另一些人则认为，随着各大公司开始构建旨在充当AI智能体(agent)的系统，蒸馏技术的重要性将会降低。训练这些智能体（即能够利用其他软件来执行任务的数字助手）的难度极大，很难通过蒸馏来复制。

Distillation “won’t matter as much for the next era of A.I.,” said Sara Hooker, chief executive of Adaption, an A.I. research lab.

AI研究实验室Adaption的首席执行官萨拉·胡克表示，蒸馏“对下一个AI时代来说将不再那么重要”。