Anthropic accuses Alibaba of ‘illicitly’ accessing AI models (opens original article in a new tab)

Anthropic accused Alibaba of using fraudulent accounts to access its AI model and criticized Chinese companies for using adversarial distillation to develop rival systems at lower costs.

• Anthropic accused Alibaba of using 25,000 fraudulent accounts to access its Claude AI model
• Anthropic claimed Alibaba's actions involved 28.8 million exchanges with Claude between April and June
• Anthropic warned that Chinese labs use adversarial distillation to develop rival chatbots without training costs
• U.S. AI companies are collaborating to combat unauthorized distillation of their models