Anthropic accuses Alibaba of extracting Claude AI capabilities in largest known distillation campaign

The accusation

Anthropic has formally accused Alibaba of conducting the largest known distillation attack against its Claude AI model. In a letter dated June 10 to Senators Tim Scott and Elizabeth Warren, the company said operators affiliated with Alibaba's Qwen AI lab used nearly 25,000 fraudulent accounts to generate more than 28.8 million exchanges with Claude between April 22 and June 5, 2026. The exchanges targeted software engineering and agentic reasoning, which Anthropic describes as Claude's most commercially valuable capabilities. Alibaba did not respond to requests for comment.

Scale and method

Distillation involves training a less capable model on the outputs of a stronger one, effectively copying its abilities. This marks the first time Anthropic has publicly named a major Chinese technology conglomerate as the source of such an attack. The Alibaba campaign alone exceeded the combined volume of three earlier campaigns attributed to Chinese startups DeepSeek, MiniMax, and Moonshot AI in February, which together generated about 16 million exchanges through roughly 24,000 fake accounts. The White House had flagged distillation as a national security concern in April, when OSTP Director Michael Kratsios published a memo committing to share intelligence with US AI labs about foreign campaigns. Anthropic noted that the Alibaba operation took place after that warning, in defiance of the administration's stance. Elon Musk recently admitted in court that he used distillation to train his Grok chatbot on outputs from OpenAI models.

Policy demands

Anthropic used the letter, sent ahead of a Senate Banking Committee hearing on AI, to urge Congress to take tougher action against Chinese companies that use distillation. It also called for allowing American AI firms to warn each other about such attacks and for further restricting Chinese access to AI chips. This stance puts Anthropic at odds with Nvidia, which has lobbied the US government to permit exports of more powerful AI systems to China.

Regulatory pressure on Anthropic

The accusation comes as Anthropic itself faces regulatory headwinds. The US government recently forced the company to temporarily withdraw its top AI software, known as Fable 5, from the market. Officials cited alleged vulnerabilities that could allow the software's built-in restrictions to be bypassed, raising concerns about potential misuse for cyberattacks. The software's ability to search for security flaws in other programs is blocked. Anthropic has argued the issues are not serious, and negotiations with the government are ongoing.

Market and geopolitical fallout

Alibaba's American depositary receipts fell more than 3% on Wednesday, dropping below $100. The decline adds to pressure on the company, which was added to the Pentagon's Chinese military companies blacklist on June 8. Alibaba sued the Defense Department this week to be removed from that list, calling the designation baseless. The distillation accusation now opens a second front in Washington.