Anthropic Accuses Alibaba of Massive AI Model Theft

In a formal letter sent to US senators and White House officials, the AI developer detailed how the operation targeted specific capabilities, including software engineering and agentic reasoning. By flooding Claude with technical queries, the perpetrators bypassed the billions in research and development expenses usually required to reach frontier-level intelligence. Anthropic confirmed it has purged all involved accounts from its platform and is now lobbying Congress to impose sanctions on firms that engage in systematic model harvesting.

Alibaba declined to comment on the allegations. The incident highlights the growing urgency for security cooperation within the industry, as Anthropic, OpenAI, and Google have already established a joint intelligence-sharing network to detect and block similar extraction efforts. The breach underscores the vulnerability of large language models to high-volume automated scraping designed to reverse-engineer competitive advantages.