Skip to main content
← Back to sources

Anthropic Discloses Industrial-Scale Distillation Campaign: 16M Claude Queries from Chinese AI Firms

Published 2026-02-23AI Regulation and GovernanceHigh

Summary

On February 23, 2026, Anthropic disclosed that three Chinese AI companies — DeepSeek, Moonshot AI, and MiniMax — conducted industrial-scale distillation campaigns against Claude models using over 24,000 fake accounts and 16 million API queries. The campaigns systematically extracted reasoning, agentic, and coding capabilities to train competing models. This disclosure followed OpenAI's February 12 warning to Congress about similar DeepSeek activity against its models. ### Scale and Targeting by

Alignment: New signal not yet covered
Related Positions: ai-governance-and-risk.md
Related Partnerships: anthropic-claude.md
anthropicclaudedeepseekminimaxmoonshotchinadistillationmodel-theftapi-securityexport-controlsnational-securityfake-accountsgeofencingcongressopenai
Anthropic Discloses Industrial-Scale Distillation Campaign: 16M Claude Queries from Chinese AI Firms — Intelligence — Agentic Developer Tools Radar · Signal