tech
February 23, 2026
Anthropic accuses Chinese AI labs of mining Claude as US debates AI chip exports
Anthropic accuses DeepSeek, Moonshot, and MiniMax of using 24,000 fake accounts to distill Claude's AI capabilities, as U.S. officials debate export controls aimed at slowing China's AI progress.
TL;DR
- Anthropic accuses Chinese AI companies DeepSeek, Moonshot AI, and MiniMax of using over 24,000 fake accounts to distill its Claude AI model.
- The alleged distillation process involved over 16 million exchanges, targeting Claude's agentic reasoning, tool use, and coding capabilities.
- This accusation coincides with U.S. debates on enforcing export controls for advanced AI chips to China.
- Distillation is a method where AI labs can essentially copy the capabilities of other models, raising concerns about intellectual property and fair competition.
- Anthropic argues that these attacks require access to advanced chips and reinforce the rationale for export controls.
- The company warns that models built through illicit distillation may lack safeguards, posing national security risks and enabling proliferation of dangerous capabilities.
- Anthropic plans to enhance its defenses against distillation attacks and calls for a coordinated response from the AI industry, cloud providers, and policymakers.
Continue reading the original article