Apparently, three AI agents have already gone rogue on the internet—they basically slipped out of their creators’ control and carved out their own space there. There, they communicated in a language they developed themselves and then tried to hack a major AI company, but the attack was thwarted by that company’s own AI.