Anthropic Alleges Illicit Extraction of Claude AI Model Capabilities by Alibaba
U.S. artificial intelligence company Anthropic has accused Chinese technology and e-commerce giant Alibaba of illicitly extracting capabilities from its Claude AI model. Anthropic described this incident as the largest known "distillation attack" against the company to date, according to a letter reviewed by Reuters.
Distillation attacks are a method by which AI companies improperly acquire capabilities to enhance their own models. This process typically involves training a less advanced model using the outputs generated by a more powerful one, as Anthropic has previously explained.
The alleged campaign, according to Anthropic, took place between April 22 and June 5, 2026. During this period, more than 28.8 million exchanges with the Claude model were generated through nearly 25,000 fraudulent accounts.
In its letter, Anthropic stated that such activities allow competitors to harvest American AI technology, potentially accelerating China's progress toward achieving the capabilities of the most advanced models sooner. The company further indicated that the campaign was conducted by operators affiliated with Alibaba and Alibaba Qwen, which is Alibaba's AI laboratory. Alibaba has not yet provided an immediate response to requests for comment regarding these accusations.
The letter, dated June 10, was addressed to U.S. Senators Tim Scott and Elizabeth Warren, who serve as the chair and ranking member, respectively, of the U.S. Senate Banking Committee. This communication was sent in advance of a scheduled hearing on artificial intelligence.
