Anthropic tells US Senate: Alibaba ran 28.8M fake Claude exchanges to distill Qwen
Anthropic's June 10 Senate letter accuses Alibaba of running 28.8 million exchanges through roughly 25,000 fraudulent accounts between April 22 and June 5 — the largest distillation attack the company has disclosed — to extract Claude's capabilities and train Qwen. The case is now a legislative front: the Trump administration had already warned against such campaigns, Anthropic is pushing Congress to mandate industry-government coordination against industrial-scale model theft, and a concurrent export control directive blocking foreign nationals from Claude's latest models signals how far the US government is willing to go to contain the leak.
Source: cnbc.com ↗
We believe combating the threat of illicit distillation requires coordinated action between government and industry, and we will continue working with Congress and the Administration to maintain American AI leadership.
Why this matters
- → 28.8M fraudulent exchanges show industrial-scale model theft now threatening U.S. AI competitiveness
- → Exposes gap between government warnings and enforcement; Alibaba proceeded despite Trump admin warnings
- → Triggers export controls blocking foreign access to latest Claude, reshaping AI industry access