ai
Anthropic Accuses Alibaba of Extracting Claude AI Capabilities in June Letter
Anthropic detailed nearly 29 million exchanges with its Claude model through fraudulent accounts linked to Alibaba. The June 10 letter to two U.S. senators called the effort the largest distillation attack the company has recorded.
UN
geeky-gadgets.comAudio version
Tap play to generate a narrated version.
Developing·Limited corroboration so far. This page will refresh as more sources emerge.
Anthropic sent a letter dated June 10 to U.S. Senators Tim Scott and Elizabeth Warren accusing operators linked to Alibaba of carrying out the largest known campaign to extract capabilities from its Claude AI model. The San Francisco company said the operators used thousands of fraudulent accounts to conduct almost 29 million exchanges.
The attacks focused on Claude's handling of longer and more complex tasks and its decision-making processes. Anthropic stated that the campaign occurred on an industrial scale to harvest and repackage U.S. AI capabilities.
It urged Congress to penalize companies behind such attacks and to strengthen measures against theft of American technology. The letter said distillation attacks convert hundreds of billions of dollars in U.S. investment and research into a subsidy for geopolitical competitors.
It referenced U.S. Department of Defense claims that Alibaba, BYD, and Baidu maintain ties to the Chinese military. Alibaba sued the U.S. government this week to seek removal from the Pentagon blacklist.
The company is listed on the New York Stock Exchange. Anthropic and OpenAI are both preparing for stock-market debuts. OpenAI and other U.S. developers have previously accused Chinese groups of using distillation attacks to train rival models at lower cost.
Some of Anthropic's advanced models, such as Mythos, have raised separate cybersecurity concerns over their ability to identify weaknesses in computer systems.
