Anthropic Accuses Alibaba of AI Capability Theft, Cites 'Largest Known Distillation Attack'
AI safety firm details alleged illicit extraction of its technology by Chinese tech giant Alibaba in a letter to Congress.
Artificial intelligence firm Anthropic has accused Chinese technology giant Alibaba of "brazenly" and "illicitly" attempting to extract its AI capabilities, describing the actions as "the largest known distillation attack on Anthropic to date." The accusation was detailed in a letter sent by Anthropic to the Senate Committee on Banking, Housing and Urban Affairs.
Distillation is an AI training method where a smaller, less advanced model is developed using the outputs of a more sophisticated model. According to Anthropic, operators affiliated with Alibaba and its Qwen AI lab conducted approximately 28.8 million exchanges with Anthropic's Claude models between April 22 and June 5. This was allegedly carried out using around 25,000 fake accounts.
"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," an Anthropic spokesperson stated. The company also claimed that Alibaba "ignored the Trump Administration’s warnings" regarding such practices.
Anthropic's letter was addressed to Senators Tim Scott and Elizabeth Warren. It comes less than two months after the White House Office of Science and Technology Policy issued a memorandum aimed at assisting AI companies in detecting and coordinating against mass distillation campaigns. This incident highlights concerns about the growing influence of powerful and affordable Chinese AI models potentially challenging America's lead in the field.
This is not the first time Anthropic has raised concerns about industrial-scale distillation. In February, the company reported identifying three similar campaigns from Chinese AI labs: DeepSeek, Moonshot, and MiniMax. Anthropic has characterized these campaigns as increasingly sophisticated and intense, urging policymakers and tech firms, including cloud providers, to collaborate on thwarting future attacks.