Anthropic claims Alibaba-affiliated operators used 25,000 fraudulent accounts to distill its AI model, urging U.S. lawmakers to tighten controls.
Anthropic accused Alibaba-affiliated operators of conducting the largest known AI model distillation campaign, extracting 28.8 million exchanges with its Claude chatbot between April 22 and June 5. The company said nearly 25,000 fraudulent accounts targeted advanced capabilities like agentic reasoning and software engineering, allowing rivals to replicate frontier AI behavior without training costs.
The alleged attack prompted Anthropic to urge Congress to strengthen export controls and penalize firms engaged in large-scale model extraction. Lawmakers are currently considering legislation to address unauthorized access to U.S. AI models. Anthropic’s letter to Senate leaders highlighted the campaign’s scale and brazen nature, noting Alibaba’s U.S. business ties.
Anthropic did not disclose immediate market reactions, but the call for regulatory action could influence policy discussions on AI security and export restrictions.