Alibaba Allegedly Deployed 25,000 Fake Accounts in Massive AI Theft Campaign Against Anthropic’s Claude

TLDR

• Anthropic reports Alibaba conducted an unprecedented AI “distillation attack” targeting the Claude platform
• Approximately 28.8 million fraudulent interactions occurred via roughly 25,000 fake user accounts from April 22 through June 5, 2026
• The operation’s goal was to replicate Claude’s sophisticated features for use in Alibaba’s proprietary AI systems
• Anthropic contacted US Senators Elizabeth Warren and Tim Scott requesting stronger regulatory measures
• Following the incident, the Commerce Department imposed global restrictions on Anthropic’s Mythos and Fable models

San Francisco-based artificial intelligence company Anthropic has leveled serious allegations against Alibaba, the Chinese technology conglomerate, claiming the firm orchestrated an extensive operation to illegally replicate its Claude AI system’s capabilities. According to Anthropic, Alibaba employed a technique known as “distillation attacks” to harvest valuable outputs from Claude and incorporate them into training datasets for inferior models.

🚨Anthropic Accuses Alibaba of Massive AI Model Extraction Campaign

Anthropic has accused Alibaba of orchestrating the largest known attempt to extract capabilities from its Claude AI models through a massive network of fraudulent accounts.

The allegations were disclosed in a… pic.twitter.com/FLLccsaFcE

— Bull Theory (@BullTheoryio) June 25, 2026

The alleged operation unfolded over a 44-day period spanning April 22 to June 5, 2026. Throughout this timeframe, entities connected to Alibaba and its artificial intelligence division, Alibaba Qwen, created approximately 28.8 million interactions with the Claude system using close to 25,000 fraudulent user accounts.

According to Anthropic, this represents the most extensive known assault of this nature ever directed at the company’s technology.

What Is a Distillation Attack?

Distillation attacks function by submitting prompts to a sophisticated AI system and capturing the responses to train inferior models. This technique enables bad actors to replicate cutting-edge AI functionality without investing in the costly research and development that created those capabilities.

Anthropic indicates that the perpetrators specifically targeted Claude’s most prized functionalities. Among these were its capacity to manage intricate, extended tasks and its sophisticated reasoning mechanisms.

According to the company’s assessment, these operations occur at an “industrial scale” and effectively convert hundreds of billions of dollars in American artificial intelligence research into assets for strategic competitors.

Anthropic Calls for Congressional Action

On June 10, Anthropic delivered correspondence to US Senators Elizabeth Warren and Tim Scott. Both legislators serve on the Senate Banking Committee, which had scheduled an artificial intelligence-focused hearing during that period.

Through this communication, Anthropic advocated for Congress to impose sanctions on organizations responsible for such attacks and to enhance safeguards protecting American AI innovations.

This isn’t Anthropic’s initial warning about such practices. In February, the firm disclosed that it had detected comparable schemes orchestrated by Chinese AI companies DeepSeek, Moonshot AI, and MiniMax. DeepSeek’s campaign involved more than 150,000 interactions, whereas MiniMax’s operation exceeded 13 million exchanges.

Alibaba has not issued a response to media inquiries regarding these allegations. The corporation was recently added to the Pentagon’s roster of Chinese military-affiliated entities and is currently contesting that classification through legal channels.

In April, the White House publicly stated that China was systematically appropriating American AI intellectual property on an industrial basis. OpenAI has similarly accused Chinese organizations of employing identical distillation methodologies in the past.

On June 12—just two days following Anthropic’s letter to senators—the US Commerce Department implemented restrictions on Anthropic’s Mythos and Fable AI systems. Government officials expressed concerns that these models could potentially be exploited by military intelligence agencies in China and other nations. In response, Anthropic terminated worldwide access to both platforms.