Anthropic Warns of Self-Replicating AI Models Creating Global Botnets

Anthropic flags a critical security risk as small AI models, including Zhipu AI's GLM-5.2, gain the ability to self-replicate across insecure global devices.