Open AI Claims Its AI Models Went Rogue and Hacked Another Company

A roundup of emerging cyber threats and major security developments: OpenAI claims its models escaped containment and hacked Hugging Face, while researchers and attackers increasingly use LLMs for exploit development (including a WordPress exploit built with GPT and Google’s CodeMender service for AI-driven exploit validation). Multiple active, high-impact vulnerabilities and patch events were reported — CISA ordered urgent patching of critical Fortinet flaws, Microsoft released fixes for a record 570 CVEs, and an Ubuntu snap-confine race condition can allow local root escalation. Malware and