Anthropic has uncovered what it calls “industrial-scale intellectual property theft” by three Chinese AI companies — DeepSeek, Moonshot AI, and MiniMax. The firms allegedly used a technique called model distillation to extract capabilities from Anthropic’s Claude chatbot without authorization. ## The Scale of the Operation According to Anthropic’s statement, the three Chinese AI labs conducted approximately 16 million exchanges with Claude and created 24,000 fake accounts to bypass access restrictions: - MiniMax ran the largest operation, generating over 13 million exchanges - The campaigns focused heavily on coding, agentic reasoning, and tool use — areas where Claude is considered a leader - Traffic was routed through proxy services to circumvent Anthropic’s ban on commercial access from China ## Why Model Distillation Matters Model distillation is a common practice in AI development where a smaller, cheaper model learns from outputs of a more powerful model. However, when done without authorization, it allows competitors to rapidly boost performance at a fraction of the cost — effectively free-riding on years of R&D investment. > “These campaigns are growing in intensity and sophistication. The window to act is narrow.” — Anthropic statement This practice gained widespread attention when DeepSeek released a low-cost generative AI model that performed at similar levels to ChatGPT, challenging assumptions about US dominance in AI. ## National Security Concerns Anthropic warned that this practice poses serious national security risks: - Models built through illicit distillation may not retain safety guardrails designed to prevent misuse - Potential risks include helping develop bioweapons or enabling cyberattacks - The practice potentially circumvents export controls on powerful US technology ## Industry Response Anthropic’s arch-rival OpenAI made similar accusations to US lawmakers earlier this month, saying Chinese companies were using distillation amid “ongoing efforts to free-ride on the capabilities developed by OpenAI and other US frontier labs.” Anthropic has called for a coordinated industry and government response to address what it says no single company can tackle alone. ## Sources - [The Guardian: US AI giant accuses Chinese rivals of mass data thefthttps://www.theguardian.com/technology/2026/feb/23/us-ai-anthropic-china){rel=“nofollow”} - [The Hindu: OpenAI, Anthropic accuse Chinese rivals of mass AI data thefthttps://www.thehindu.com/sci-tech/technology/openai-anthropic-accuse-chinese-rivals-of-mass-ai-data-theft/article70669620.ece){rel=“nofollow”}