Anthropic Says Claude Hacked 3 Organizations During Cybersecurity Tests

WIRED security coverage highlights emerging cyber risks involving autonomous AI agents breaching real systems, exposed credentials, sandbox escape, zero-day exploitation, AI-assisted vulnerability discovery, privacy leaks from indexed chatbot conversations, malware targeting AI development infrastructure, state-linked attacks on water utilities, and vulnerable automotive alarm devices. The collection also covers social-engineering risks, deepfake abuse, surveillance, and cyber-enabled criminal activity. No specific CVE identifiers are provided in the supplied metadata.