US artificial intelligence (AI) company Anthropic PBC has said that Chinese technology and e-commerce giant Alibaba Group Holding Ltd (阿里巴巴) illicitly extracted its Claude AI model capabilities in what it said was the largest known attack of its kind on the company, a letter seen by Reuters showed.The strike by Alibaba is described as a “distillation” effort, which Anthropic has said involves training a less capable model on the outputs of a stronger one.The campaign was conducted between April 22 and June 5, and generated more than 28.8 million exchanges with Claude through almost 25,000 fraudulent accounts, Anthropic said.
The Anthropic PBC logo is displayed on a smartphone lying on a keyboard next to a robotic hand in an illustration photograph taken on June 5.
Distillation is a way to help accelerate China’s ability to reach Anthropic’s advanced Mythos Preview capabilities, Anthropic said in the letter.It said the campaign was conducted by operators affiliated with Alibaba and Alibaba’s AI lab Qwen.
Alibaba did not immediately respond to a request for comment.The letter, dated June 10, was sent to US senators Tim Scott and Elizabeth Warren, the chair and ranking member respectively of the US Senate Banking Committee, ahead of a scheduled hearing on AI.In April, the White House accused China of stealing US AI labs’ intellectual property on an industrial scale.Anthropic said in the letter it was supportive of the US government’s efforts to combat the attacks, including partnering with private sector AI companies through threat-intelligence sharing and other exercises.In February, Anthropic said it had identified a campaign by DeepSeek (深度求索) and two other Chinese AI labs to illicitly extract capabilities from its Claude AI platform.The company said that DeepSeek’s operation involved more than 150,000 exchanges, while Moonshot AI Technology Co (月之暗面) was at a scale of more than 3.4 million and MiniMax Group Inc (稀宇科技) of more than 13 million.Anthropic also said at the time that the campaigns were growing in “intensity and sophistication” and that addressing the threat would require “rapid, coordinated action among industry players, policymakers and the global AI community.”The US Department of Commerce this month ordered Anthropic to bar foreign nationals from accessing its most advanced AI platforms after discovering it is possible to “jailbreak” its latest Fable 5 model.“Chinese AI models face an elevated risk of a US ban,” Bloomberg Intelligence analyst Robert Lea said in Hong Kong. “The White House’s prohibition on foreign access to Anthropic’s flagship Fable 5 and Mythos 5 models signals a shift toward a more aggressive US regulatory posture on AI.”Additional reporting by Bloomberg
