Anthropic Accuses Alibaba of Gaining Illicit Access to Claude

Anthropic accused Alibaba of gaining illicit access to its Claude AI model. AI player said that the Chinese e-commerce giant has set up almost 25,000 fraudulent accounts, violating the terms of service and access restrictions.

In a letter sent to the US Congress, Anthropic head of policy Sarah Heck stated that Alibaba used the accounts to generate more than 28.8 million exchanges with Claude, targeting some of the model’s most valuable capabilities such as agentic reasoning, software engineering, and long-horizon tasks. The alleged violation took place between 22 April and 5 June 2026, conducted by operators said to be affiliated with Alibaba and Alibaba Qwen, the company’s AI lab.

AI player noted that the campaign follows many of the same patterns it disclosed in February, spotlighting distillation attacks of DeepSeek, Moonshot and Minimax, which combined generated more than 16 million exchanges with Claude through 24,000 fraudulent accounts. “These distillation attacks are carried out illicitly, systematically, and at industrial scale to harvest US AI capabilities across frontier labs and repackage them as their own without incurring the training and R&D costs required to train frontier models,” stated Heck.

The company further warned that such activity turns hundreds of billions of dollars in American investment and R&D into a massive subsidy for geopolitical competitors. It called on Congress to advance measures that facilitate threat information sharing between US AI labs, close loopholes that allow China AI labs to access advanced US chips, and penalize labs from the nation responsible for distillation attacks. OpenAI has also accused Chinese labs of engaging in illicit behaviour in the past.