In a letter dated June 10 to U.S. Senators Tim Scott of South Carolina and Elizabeth Warren of Massachusetts, obtained by The Wall Street Journal, Anthropic alleged that operators linked to Alibaba’s Qwen AI lab used approximately 25,000 fake accounts to initiate nearly 29 million interactions with Claude over a six-week period beginning April 22. The company called it “the largest known distillation attack on Anthropic to date.”
Distillation attacks involve training a less capable AI model on the outputs of a more advanced one, enabling the attacker to replicate advanced capabilities without incurring the steep research and computing costs of developing a frontier model from scratch. Anthropic said the campaign “targeted some of Claude’s most valuable capabilities, such as agentic reasoning, software engineering, and long-horizon tasks.”
“Distillation attacks turn hundreds of billions of dollars in American investment and R&D into a massive subsidy for our geopolitical competitors,” the letter stated, according to copies seen by multiple news organizations.
The letter referenced an April memorandum from the Trump administration that warned foreign entities, principally based in China, were engaged in “industrial-scale campaigns” to distill U.S. frontier AI systems. Anthropic said it supports Washington’s efforts to combat such attacks but called for additional measures, including tightening chip controls and enacting legislation to penalize AI labs found to be engaging in illicit distillation.
An Anthropic spokesperson declined to comment on the letter’s specific contents but told multiple outlets: “We believe combating the threat of illicit distillation requires coordinated action between government and industry, and we will continue working with Congress and the Administration to maintain American AI leadership.”
Alibaba did not immediately respond to a request for comment.
The accusations are not the first from Anthropic. In a February blog post, the company said it had identified industrial-scale extraction campaigns by three other Chinese AI labs — DeepSeek, Moonshot, and MiniMax. OpenAI has also made similar allegations against DeepSeek over unauthorized use of its models.
The allegations come amid a broader push by the U.S. government to restrict foreign access to advanced AI. Earlier this month, the White House restricted foreign use of Anthropic’s top-tier models, and the Pentagon added Chinese firms including EV maker BYD and search engine Baidu to a list of companies with links to China’s military, according to details in the letter.