Anthropic accused Alibaba of using Claude to run almost 29 million exchanges through thousands of fraudulent accounts. The company said the effort was the largest extraction campaign of its kind and targeted Claude's longer-task performance and decision-making.
Anthropic said the activity used distillation attacks, which pull answers from a stronger AI model to train a weaker one. For readers, that means the dispute is not about one stray prompt run but about a scaled attempt to repurpose Claude's behavior into a separate model stack.
Tim Scott received the letter
Anthropic dated its letter to 10 June and sent it to US Senators Tim Scott and Elizabeth Warren. It urged US Congress to penalise companies behind attacks like this and to ramp up measures to prevent US tech from being stolen.
The company also said other alleged attacks posed a threat to the US military. It said the attacks were carried out on an industrial scale so Chinese companies could harvest and repackage US AI capabilities as their own.
Alibaba and the Pentagon blacklist
The accusation lands while Alibaba is suing the US government this week to remove its name from the Pentagon blacklist. Anthropic also said the US Department of Defense claims that Alibaba, BYD, and Baidu are tied to the Chinese military.
That collision leaves a practical question for US lawmakers and companies watching the case: whether Congress treats the allegation as a model-security problem, a trade dispute, or both. Anthropic's own quote puts the stakes in blunt terms: "Distillation attacks turn hundreds of billions of dollars in American investment and [research and development] into a massive subsidy for our geopolitical competitors".
Claude AI and future controls
Anthropic's complaint fits a wider pattern in which OpenAI has previously accused Chinese groups of using distillation attacks. Anthropic and OpenAI are both gearing up for a stock market debut, so claims about model theft now sit beside investor questions about how well US AI companies can protect the systems they sell.
The unanswered issue is the evidence Anthropic says ties the fraudulent Claude accounts to Alibaba-linked operators. Until that linkage is explained, the size of the allegation is clear, but the proof standard behind it is still the part that matters most.
