Anthropic Accuses Alibaba of ‘Brazenly’ & ‘Illicitly’ Stealing Their AI

Photo Credit: Towfiqu Barbhuiya

Anthropic has accused Chinese tech company Alibaba of “brazenly” and “illicitly” attempting to steal its artificial intelligence capabilities.

AI company and Claude chatbot creator Anthropic has issued a letter to the U.S. Senate Committee on Banking, Housing, and Urban Affairs, accusing Chinese tech company Alibaba of “brazenly” attempting to steal its artificial intelligence capabilities. Dated June 10, the letter alleged that Alibaba carried out “the largest known distillation attack on Anthropic to date.”

According to the Institute for AI Policy and Strategy, an AI distillation attack is a form of intellectual property theft in which a malicious actor uses a powerful AI model to train a smaller, competing model. By repeatedly launching queries through the “teacher” model via an API and recording the responses, attackers can use that data to build a high-performance “student” model.

Anthropic wrote that persons affiliated with Alibaba and its AI department carried out 28.8 million exchanges with its models across roughly 25,000 fraudulent accounts between April 22 and June 5.

“We believe combatting 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,” said an Anthropic spokesperson.

The company’s letter arrives on the heels of the White House Office of Science and Technology Policy’s memorandum pledging to assist AI companies against industrial-scale distillation. Anthropic announced in February that it had identified three industrial-scale distillation campaigns from smaller AI companies: DeepSeek, Moonshot, and MiniMax.

“These labs created over 24,000 fraudulent accounts and generated over 16 million exchanges with Claude, extracting its capabilities to train and improve their own models,” wrote Anthropic on X (formerly Twitter).

It also comes amid numerous music publishers, including BMG, Concord, Universal Music Publishing Group, and ABKCO, who are suing Anthropic with allegations that it stole copyrighted lyrics en masse to train its Claude models. The publishers assert that Anthropic caused more than $3 billion in damages, but the AI company argues that AI model training falls under fair use.