💼 Anthropic Accuses Alibaba of Stealing Claude Data
Anthropic has claimed that Alibaba conducted a large-scale "distillation attack" to steal the capabilities of its flagship Claude model. According to a letter to the U.S. Senate, the Chinese giant used approximately 25,000 fake accounts to generate nearly 29 million queries aimed at extracting advanced model skills, such as agentic reasoning and software engineering.
🌍 The incident highlights the growing threat to intellectual property in the AI field and stimulates discussions regarding the need for legislative regulation of "distillation attacks," where the capabilities of powerful models are used to train weaker ones without permission.
👤 This is an example of how the fight for AI leadership is moving into a phase of industrial espionage, which could lead to stricter controls over API access and cloud computing.
Source 1: https://www.resultsense.com/news/2026-06-25-anthropic-alibaba-claude-distillation/ Source 2: https://www.morningstar.com/news/dow-jones/20260625767/anthropic-claims-alibaba-ran-brazen-campaign-to-access-its-claude-ai-model