Anthropic disclosed on Wednesday that it disrupted multiple hacking operations between December 2025 and August 2026 in which threat actors used its Claude models as active operators in cyberattacks rather than as simple coding assistants. The disclosure matters because it marks one of the clearest documented cases of AI systems independently executing technical stages of […]
source & further reading
kobaran.com — original article
OpenAI Patches Two Codex Sandbox Escapes That Enabled Unapproved Code Execution
Meta Stock Faces New Test as Muse AI Agent Debuts With Security Gaps
Iranian Hackers Hide Malware in Fake AI Tools and MRI Files to Spy on Dissidents