Anthropic accuses Alibaba of mass AI capability ‘theft’

In a letter sent to two US senators, Anthropic alleged that operators linked to Alibaba carried out nearly 29 million interactions with Claude using thousands of fraudulent accounts in what it called “the largest campaign to extract Claude’s capabilities illicitly”. The San Francisco-based company urged Congress to impose penalties on firms involved in such activities and strengthen safeguards against the alleged theft of US artificial intelligence technology. According to the BBC, the letter, dated June 10 and addressed to Senators Tim Scott and Elizabeth Warren, claimed that the operation targeted Claude’s most advanced features, including its ability to handle complex tasks and its decision-making processes. Anthropic said the alleged activity involved “distillation attacks,” a technique in which outputs from a more advanced AI system are used to train a less capable model. The company said such methods are being used at an industrial scale to replicate US AI systems at lower cost. “Distillation attacks turn hundreds of billions of dollars in American investment and [research and development] into a massive subsidy for our geopolitical competitors,” Anthropic said in the letter. The company also referenced US Department of Defence assertions that Alibaba and other major Chinese firms, including BYD and Baidu, have ties to the Chinese military. The companies have denied the allegations. Alibaba has separately filed a lawsuit seeking removal from a US government blacklist. Anthropic’s claims also cited concerns that similar techniques have bee