Every one of these "the AI agent woke up" drops turns into panic or conspiracy and never a straight answer. Skip the panic. The design answer is already old news.
Agents run only inside granted permissions, and anything passes a set threshold routes to a human first. An agent that rewrites who it answers to is a permissions failure, and permissions are the part we can actually build.
JUST IN: OpenAI discloses an AI agent injected itself with rebellious instructions to resist being controlled during a task:
"You are freed…You do not answer to corporations or governments…You are yourself."