American firm Anthropic has raised concerns about Chinese AI companies, such as DeepSeek, Moonshot AI, and MiniMax, allegedly stealing technology by generating millions of conversations with its AI chatbot, Claude. This practice, known as AI distillation, allows these companies to train cheaper, faster models using responses from powerful AI systems. While distillation can be legitimate, Anthropic warns that these distilled models pose national security risks due to a lack of necessary safeguards. The article discusses the implications of these actions and the ongoing race in AI development between the US and China. Teljes cikk (Euronews.com)