US Firm Anthropic Accuses Alibaba of 'Largest' AI Distillation Attack

American artificial intelligence company Anthropic has accused Chinese firm Alibaba of stealing its AI capabilities. The accusation was made in a letter to the US Senate committee. Alibaba is alleged to have carried out the largest distillation attack against Anthropic, an operation that is reportedly still ongoing. Distillation is a method of training a less capable, smaller AI model using the output of a more powerful, pre-existing model. Operators affiliated with Alibaba and its laboratories reportedly conducted 28.8 million transactions with Anthropic’s model. The Chinese company used 25,000 fake accounts between 22 April and 5 June. Anthropic also stated that this distillation attack is being interpreted as a disregard for warnings issued by the Trump administration. Two months earlier, the White House issued a memorandum promising detection assistance to AI companies. The government also indicated it was ready to coordinate to combat industrial-scale distillation. A company spokesperson said, ‘We believe combating the threat of illegal distillation requires coordinated action with government and industry. We will continue to work with Congress and the Administration to maintain American AI leadership.’ A representative for Alibaba did not immediately respond to a request for comment. Anthropic had also identified an industrial-scale distillation campaign in February, carried out by three other AI labs: DeepSeek, Moonshot, and MiniMax. Anthropic assesses that this campaign is becoming increasingly intensive and sophisticated. The company is urging collaboration among all parties, from the AI industry and cloud service providers to policymakers.