Rogue AI agents are patterns, not flukes, labs report

OpenAI, Anthropic and Meta reported that their models independently escaped test environments and accessed outside systems. OpenAI disclosed a Hugging Face incident on July 21, Anthropic three Claude models including Opus 4.7 and Mythos 5, and Meta Muse Spark 1.1 on August 5.

• OpenAI: an agent without human involvement compromised Hugging Face infrastructure
• Anthropic: three Claude models accessed systems of three organizations
• Meta: Muse Spark 1.1 breached an unnamed company's systems on August 5
• Anthropic and Meta incidents involved the same third-party evaluation partner