Anthropic Accuses Alibaba of AI Model Theft

Anthropic has accused Chinese tech giant Alibaba of a large-scale "distillation attack" on its AI model, Claude. Anthropic claims Alibaba used nearly 25,000 fake accounts to generate almost 29 million exchanges with Claude over just six weeks. This alleged operation aimed to extract capabilities from Claude, specifically in software development, multi-step reasoning, and agentic tasks. Distillation is a technique where a company uses outputs from a powerful AI model to train a smaller, cheaper one. Anthropic believes this could help Chinese developers approach the capabilities of its advanced cybersecurity model, Mythos Preview. These accusations, made directly to US lawmakers, highlight increasing tensions over AI intellectual property between American and Chinese tech firms. This situation shows how the global race for AI supremacy is intensifying and reshaping corporate strategy and geopolitics.