Anthropic Accuses Alibaba of Massive Claude Cloning Operation

Anthropic has accused Alibaba of orchestrating the largest known attempt to clone Claude, allegedly generating 28.8 million interactions through nearly 25,000 fraudulent accounts between April and June. According to Anthropic, the campaign targeted Claude’s most valuable capabilities, including agentic reasoning, software engineering and long‑horizon tasks, using obfuscation techniques to evade detection. The company claims Alibaba sought to extract Claude’s abilities without paying the massive R&D costs required to build a frontier model, despite US warnings against such distillation attacks. Anthropic is urging Congress to impose penalties, expand export controls and update antitrust rules to allow AI firms to share intelligence on Chinese tactics. The company warns that unchecked cloning efforts could accelerate China’s AI progress and undermine US national security, especially as Chinese leaders openly describe US frontier models like Mythos as strategic cyber weapons.

Read the full story on Ars Technica →