analiza aiqbench

Anthropic alleges that Chinese rivals DeepSeek and Moonshot covertly routed user queries through its Claude models to refine their own systems, marking a significant escalation from previous training data disputes. This direct re-routing practice raises critical concerns regarding data privacy, intellectual property theft, and the geopolitical tensions surrounding cross-border AI infrastructure usage. The situation highlights emerging risks in model deployment strategies where competitors may exploit proprietary APIs without explicit consent.