TLDR Tech / InfoSec
Anthropic Accuses Alibaba of Claude Distillation/Cloning Attack
Anthropic alleges that Alibaba and its Qwen team orchestrated a large-scale AI model distillation campaign using approximately 25,000 fake accounts to conduct 28.8 million interactions with Claude from April 22 to June 5, leveraging obfuscation and proxy networks to extract capabilities in agentic reasoning, software engineering, and long-horizon tasks. The company has reported the incident to US Senate and White House officials, warning that such theft could significantly accelerate China's AI advancement by cloning proprietary model behaviors. Anthropic urges Congress to reform antitrust laws and advises enterprises to strengthen monitoring for AI model exploitation and data leakage vectors.
Read the full story