Anthropic identifies 24,000 fraudulent accounts extracting Claude's capabilities

Anthropic says it has detected industrial-scale “distillation attacks” on its Claude models, carried out by DeepSeek, Moonshot AI and MiniMax. The company says the labs created over 24,000 fraudulent accounts and generated more than 16 million exchanges with Claude, thereby “extracting its capabilities to train and improve their own models.” While acknowledging that “distillation can be legitimate,” Anthropic warned that “foreign labs that illicitly distil American models can remove safeguards, feeding model capabilities into their own military, intelligence, and surveillance systems.” The company added that the attacks are “growing in intensity and sophistication” and called for “rapid, coordinated action among industry players, policymakers, and the broader AI community.” Anthropic says it attributed each campaign “to a specific lab with high confidence through IP address correlation, request metadata, infrastructure indicators, and in some cases corroboration from industry partners.” X user AntonLaVay pointed out this claim suggest

anthropic, claude, distillation attacks, deepseek, moonshot ai, minimax, fraudulent accounts, model extraction, ip correlation, model safeguards