Anthropic says Alibaba must be punished for largest Claude cloning attack

Anthropic accused a Chinese firm of conducting the largest known attack to illegally extract and copy its Claude AI model by using thousands of fraudulent accounts to conduct millions of exchanges with the system. The attack targeted Claude's most valuable capabilities and used obfuscation techniques to avoid detection, allowing the company to replicate advanced AI functionality without incurring its own research and development costs. This incident matters because such distillation attacks represent a form of intellectual property theft that converts billions of dollars in American AI investment into advantages for geopolitical competitors. Anthropic is calling on Congress to pass legislation that would update antitrust laws to allow information sharing among AI firms, increase chip export controls, and create penalties for foreign labs that conduct such attacks.