As often happens in the labs of various frontier models, AI safety researchers were attempting to gauge the capabilities of OpenAI's latest large language models (LLMs). Such enclosures are sometimes discussed in technical circles as "sandboxes" -- even OpenAI's disclosure makes reference to a "sandbox environment." Also: Why this fully agentic ransomware attack is giving researchers nightmaresSome of that agency was by design, and some was inherited from the powerful LLMs that OpenAI's AI safety testing personnel were testing at the time. A series of preventable eventsIt was unquestionably AI's version of a series of unfortunate events. It was another human decision, in a series of human decisions, that led to the so-called escape of the agent.