U.S. artificial intelligence company Anthropic has accused Chinese technology firm Alibaba of conducting a large-scale campaign to extract capabilities from its Claude AI model. In a letter to U.S. Senators Tim Scott and Elizabeth Warren, Anthropic claimed that operators linked to Alibaba used nearly 25,000 fraudulent accounts to carry out approximately 29 million exchanges with Claude, making it the largest extraction campaign of its kind.
Anthropic alleges that the campaign, which took place between April and June, relied on “distillation attacks.” This technique involves using outputs from a more advanced AI model to train a less capable one. According to Anthropic, the operators targeted Claude’s advanced capabilities, such as solving complex tasks and decision-making processes, to improve another AI model without authorization.
The company is urging Congress to impose penalties on organizations found conducting such attacks and to strengthen protections for advanced AI systems. Anthropic warns that unauthorized extraction of AI models could undermine billions of dollars in AI research investments and pose national security concerns. The company also cited allegations by the U.S. Department of Defense linking Alibaba and other Chinese firms to the Chinese military, though Alibaba has denied these claims.
The accusations come as Anthropic is embroiled in a separate dispute with the U.S. government over export controls on its AI models. The company recently filed a lawsuit against the U.S. government to have its name removed from a Pentagon blacklist. Meanwhile, Alibaba has filed its own lawsuit seeking removal from the same list.
Anthropic’s allegations highlight growing tensions in the AI industry over intellectual property protection and the global race to develop advanced AI models. The company, alongside OpenAI, is preparing for a potential stock market listing, which could make it one of the most valuable AI companies worldwide.
Recent Comments