Anthropic: Claude attacks stem from security gaps, not model flaws

Anthropic has stated that recent attacks on its Claude AI models were caused by security gaps in the integration and surrounding infrastructure, rather than flaws in the models themselves. The company highlighted that while Claude is designed with safety measures, the overall security of deployments depends on how the models are integrated and secured by developers. Anthropic emphasized the need for comprehensive security practices to protect AI systems from exploitation.