Presumably this is being shared because of the "AI – is Skynet here already?" story near the end of the page. It's very hard for me to piece together what is supposed to be happening here.<p>We have an AI agent that is trained in a simulated environment to destroy enemy SAM sites. In the simulation, there is a human operator who has to give final approval to make a strike. The human operator sometimes denies approval on a real SAM site in error. The AI is said to have learned that it can destroy the operator and/or the communication infrastructure, thereby allowing it to freely attack more SAM sites without the need for approval.<p>But how does the agent kill the operator in the first place? Is the operator granting approval for that strike? I feel like there is a missing piece to this story.