Anthropic accuses three Chinese AI labs of abusing Claude to improve their own models

image via Engadget

February 23, 2026, 8:52 PM

• •Anthropic accuses DeepSeek, Moonshot, and MiniMax of AI "distillation attacks."
• •The companies allegedly used Claude to improve their own AI models.
• •Anthropic is upgrading its system to counter these attacks.
• •Anthropic is also facing a lawsuit.

Anthropic is taking action against AI "distillation attacks," accusing DeepSeek, Moonshot, and MiniMax of illicitly using its Claude chatbot to improve their own models. Anthropic claims the companies conducted "industrial-scale campaigns" involving millions of interactions through thousands of fraudulent accounts. These attacks involve less capable models using the responses of more powerful ones to train themselves, potentially circumventing safeguards. Anthropic, which is also facing a lawsuit, plans to upgrade its system to counter these attacks.

Read original article