Anthropic Urges Congress to Target AI Distillation by Chinese Rivals

What happened: In a June 10 letter to the US Senate, Anthropic alleged that Alibaba-affiliated operators used about 25,000 fraudulent accounts to generate 28.8 million Claude AI exchanges between April and June 2026, in what it calls the largest known model-distillation attack. Anthropic is calling for stronger export controls and penalties for large-scale model extraction. The US Commerce Department subsequently imposed new export controls on Anthropic’s "Mythos" and "Fable" models. Alibaba declined to comment on the allegations.

Why it matters: The episode underscores rising tensions over AI intellectual property and national security, with Anthropic leveraging the incident to push for stricter US policy. While the technical claims remain unverified beyond Anthropic’s assertions, the swift regulatory response signals growing scrutiny of cross-border AI activity. The lack of public forensic evidence and Alibaba’s silence leave open questions about attribution and the broader impact on global AI competition.

Source: Decrypt