The clash is between titans: the American artificial intelligence company anthropic accused Alibaba, the Chinese technology and e-commerce giant, to have stolen illegally some features of its artificial intelligence model Claude in what he defined the largest attack of its kind never happened against the company, according to a letter seen by ReutersThe action taken by Alibaba, which lost 4% on the Hong Kong stock exchange, is described as a “distillation” operation, which according to Anthropic technically consists of training a less performing model on the results of a more powerful model, but in reality it is above all a way to To make grow the ability of China to reach the advanced features of Mythos Anthropic Preview.
Anthropic said the operation was conducted between April 22 and June 5, 2026, and generated over 28,8 million trades with Claude through nearly 25.000 fraudulent accounts. It was conducted by operators affiliated with Alibaba and Alibaba Qwen, Alibaba's artificial intelligence lab.
The letter, dated June 10, was sent to the senators Tim Scott and Elizabeth Warren, respectively chairman and minority member of the U.S. Senate Banking Committee, ahead of a scheduled hearing on artificial intelligence. In April, the White House ha China accused of having stolen the intellectual property of US artificial intelligence laboratories on an industrial scale.
In a post last February, Anthropic had already identified a campaign conducted by the Chinese artificial intelligence startup DeepSeek, whose low-cost AI model rocked the tech world in January 2025, and two other Chinese AI labs, with the intention of illegally stealing functionality from its Claude AI platform: DeepSeek's operation involved over 150.000 trades, while Moonshot AI reached a scale of over 3,4 million and MiniMax over 13 million. At the time, it was also stated that the campaigns were becoming increasingly "intense and sophisticated" and that addressing the threat would require "rapid and coordinated action among industry players, policymakers, and the global AI community."
Alibaba Listed by Pentagon as a Military Company
This month Alibaba has been added to the list of Pentagon with the Chinese military companies. Instead, the Department of Commerce avoided including DeepSeek in a commercial blacklist, as reported by Reuters this month, despite being deemed a threat to national security by a government committee, in an attempt to avoid escalation tensions with Beijing.
Meanwhile, on June 12, two days after Anthropic sent the letter, the Department of Commerce imposed some controversial restrizioni on Anthropic's latest AI models Mythos and Fable, as officials fear they could be used by military intelligence in China and other countries deemed at risk. The restrictions have led Anthropic to disable access to the models globally.
