Anthropic details distillation attack campaigns from China (Alibaba, Moonshot AI, and DeepSeek)

Bad news! Serious stuff! The Communist Party of China is well known for its intellectual property theft and espionnage of the West. Competing by illegal means! Very troubling!

What seems to be particularly troubling is that these Chinese companies are apparently sharing the illegal means facilitating these distillation and other attacks.

"A new report released Thursday by Anthropic alleged persistent distillation attacks by China-based AI companies, which have escalated in recent months as competition in the space has intensified.

“Over the last several months, unauthorized labs have developed increasingly sophisticated methods to circumvent our defenses and harvest the capabilities of US frontier models,” the report reads. “The campaigns we identified targeted some of Claude’s most valuable capabilities, including agentic capabilities and tool use, coding and data analysis, and logical reasoning.”

Anthropic previously spoke out about distillation attacks in February, even calling out specific labs.

OpenAI has reported similar activity, which it attributed to DeepSeek specifically. But the campaigns detailed in Anthropic’s new report are both larger and more aggressive. All told, the company observed nearly 200 million exchanges linked to distillation attacks, attributed to five separate campaigns. ..."

"... We define illicit distillation as an industrial-scale, covert campaign to extract a model’s capabilities and replicate them in another model without authorization. Illicit distillation is typically enabled by fraud: sophisticated networks of fake accounts created with stolen credit cards, login credentials, and API keys. ...

Beyond distillation, Alibaba also used Claude to advance its AI R&D efforts. Alibaba used Claude to help develop its internal infrastructure for model development. Claude was used to help develop Alibaba’s reinforcement learning (RL) environments and advance model architecture research.

Alibaba accessed Claude through two main pools of fraudulent accounts. The first consisted of nearly 5,000 fraudulent accounts leveraging residential proxies, disposable emails, and virtual-card payments to obfuscate their access.

When we banned this pool of accounts, Alibaba quickly shifted its traffic through the second pool.

Some of these accounts were found to have been funneling requests from DeepSeek and Xiaomi, demonstrating that the same proxy service networks are often used by a variety of organizations. ..."

Anthropic details distillation campaigns from Alibaba, Moonshot AI, and DeepSeek | TechCrunch

Detecting and countering misuse of AI: September 2026

The graphic below illustrates the life cycle of an illicit distillation campaign.

Anthropic sued (again) for training on song lyrics and US government sides with OpenAI on issue of training LLMs on copyrighted material

What is the training of large models on copyrighted material (intellectual property)? Fair use or pirating? This answer to this question becomes very urgent.

The statutory damages asked for by plaintiffs could easily bankrupt Anthropic or any other company involved in model training.

I again strongly suggest to reform intellectual property rights. They are effective for way too many years! E.g. books and songs in the US, they "last for the life of the author plus an additional 70 years." (Google search) This is insane!!!

"In a lawsuit that The New York Times filed against OpenAI, the Trump administration has contributed a 20-page brief in defense of the ChatGPT maker’s unlicensed use of copyrighted material to train its LLMs.

“The United States has a strong interest in continuing to develop a robust and competitive artificial intelligence industry that sets the standard for the practice and procedure of AI use globally… As such, it is critical for the United States to ‘retain global leadership in artificial intelligence,’” the brief reads, referencing an executive order that President Donald Trump signed last year. ..."

"Sony Music Publishing and Warner Chappell Music sued Anthropic, along with CEO and co-founder Dario Amodei and co-founder Benjamin Mann, in the U.S. District Court for the Northern District of California, alleging Claude was trained on tens of thousands of copyrighted songs without permission. The complaint accuses Anthropic of torrenting pirated books from Library Genesis and Pirate Library Mirror, scraping lyrics from licensed sites like Musixmatch and LyricFind, and running a “destructive scanning” operation on physical books, citing findings already unsealed in the separate Bartz v. Anthropic authors’ case. The publishers seek statutory damages up to $150,000 per willfully infringed work plus up to $25,000 per removed copyright notice, destruction of infringing copies, and disclosure of Claude’s training data.
This is the fifth major music-industry suit against Anthropic, following actions from Universal Music Publishing Group, Concord, and ABKCO (over $3 billion), BMG, and Round Hill Music.
It also follows Anthropic’s $1.5 billion settlement with book authors over similar torrenting conduct, announced in 2025 and approved this year. The suit signals that music publishers are pushing for damages and licensing terms rather than accepting a one-time settlement. ..." (DataPoints)

Now Sony Music Publishing and Warner Chappell sue Anthropic in multi-billion dollar lawsuit: ‘One of the largest and most blatant ongoing thefts of intellectual property in history.’

US government sides with OpenAI on issue of training LLMs on copyrighted material | TechCrunch

US threatens sanctions against Chinese AI models over IP theft

Good news! 

I blogged here about a recent claim of a Chinese cloning attack or was it distillation.

"On Tuesday, Treasury Secretary Scott Bessent said the U.S. would examine open source models from China for signs of intellectual property theft, threatening sanctions against Chinese AI companies if IP theft is established. 

“We’ve seen a lot of talk about open source models coming and threatening the large language models in the U.S.,” Bessent said on Fox Business Tuesday. “This administration supports open source models, but what we do not support is IP theft. If we see, especially, that overseas models are stealing from our great companies, we have the ability to sanction them because of this theft.” ..."

US threatens sanctions against Chinese AI models over IP theft | TechCrunch

Taiwan alleges ex-TSMC staff stole chip secrets to sell to China

Bad news! Sounds familiar!

"Taiwanese prosecutors on Monday indicted a former TSMC deputy manager for allegedly stealing trade secrets involving national core technologies with the intent of using them to advance China's chip ambitions. ..."

Taiwan alleges ex-TSMC staff stole chip secrets to sell to China - Nikkei Asia "Indictment is first under National Security Act involving attempt to advance Chinese chip ambitions"

Anthropic says Chinese Alibaba must be punished for largest Claude cloning attack

Serious stuff! China has long been accused of allowing the stealing of intellectual property!

"Anthropic has accused the Chinese firm Alibaba of launching the largest attack yet attempting to clone Claude, as China races to match the capabilities of Anthropic’s leading model following Mythos’ release and subsequent restriction from foreign markets. ..."

Anthropic says Alibaba must be punished for largest Claude cloning attack - Ars Technica "Alibaba allegedly used 25,000 accounts to mine Claude over 28.8 million exchanges."

Anthropic accused three Chinese AI companies of setting up more than 24,000 fraudulent accounts

This is a serious allegation!

"Anthropic accused three Chinese AI companies of setting up more than 24,000 fraudulent accounts with its Claude AI model to help their own systems catch up.

The U.S. AI startup said that DeepSeek, Moonshot AI and MiniMax prompted Claude more than 16 million times to siphon data and that the activity raised national-security concerns for the U.S. Earlier this month, OpenAI, an Anthropic rival, sent a memo to House lawmakers accusing DeepSeek of using the same tactic to mimic its products. Representatives from the trio of Chinese companies didn’t respond to requests for comment."

"Anthropic is accusing three Chinese AI companies of setting up more than 24,000 fake accounts with its Claude AI model to improve their own models.

The labs — DeepSeek, Moonshot AI, and MiniMax — allegedly generated more than 16 million exchanges with Claude through those accounts using a technique called “distillation.” Anthropic said the labs “targeted Claude’s most differentiated capabilities: agentic reasoning, tool use, and coding.” ..."

The Wall Street Journal What's news

Anthropic accuses Chinese AI labs of mining Claude as US debates AI chip exports

Is Huawei Stealing the West’s Top Tech Talent? with Palki Sharma

Recommendable!

From stealing IP and imitation to poaching! I blogged here earlier about this subject. Western countries, Taiwan, South Korea are already responding to this.

China’s latest Weapon to Nab Western Technology—Through Its Courts

Very concerning! Very clever by the Communist Party of China!

Courts in China is an oxymoron! They are Kangaroo courts!

"China’s courts are invalidating western tech patents" (Source)

China’s Newest Weapon to Nab Western Technology—Its Courts - WSJ Rulings nullify patents in industries it deems important, including technology, pharmaceuticals and rare-earth minerals

The Wall Street Journal  apparently has been tracking this issue for some time:

How China Systematically Pries Technology From U.S. Companies (dated 9/26/2018) Beijing leans on an array of levers to extract intellectual property—sometimes coercively

How China Spies on Countries Throughout Western Hemisphere

Recommendable! The following article probably only scratches the surface of this serious issue!

"A federal court in Virginia recently granted Microsoft the authority to seize websites used by a Chinese state-linked hacking group.

Known as Nickel, the group has successfully committed cyberattacks in up to 29 countries, including the United States,and 16 Latin American and Caribbean nations. ...
Much of Latin America and the Caribbean is seen as a recurring target, given prolonged vulnerabilities in cyber infrastructure and a shortage of cybersecurity professionals. ...
The report also highlights repeated accusations against Chinese government entities of cyberespionage in the region favoring Chinese companies, often state-owned or state affiliated. Cyber-intelligence enables access to critical market information, coercion of local officials, intellectual-property theft against competitors, and influence operations aimed at decreasing local resistance to China. ..."

How China Spies on Countries Throughout Western Hemisphere

From China, with Spies | Rob Schmitt Tonight

Very recommendable! Very concerning!

Massive intellectual property theft in the 300-600 billions of dollars. Significant funding of BLM and other divisive propaganda! Collecting enormous amounts of information including DNA of basically all U.S. citizens. Targeting prospective politicians to influence or coerce their future careers.

The FBI has about 2,000 open cases on Chinese espionage and influence operations! Every 10 hours, the FBI opens up a new investigation into such cases.

Mark Warner On CCP Threat: 'This Story Needs To Get Out To The American Public'b4