If the world is to be filled with innumerable discrete human level intelligences, the most plausible reason I can imagine them all secretly and flawlessly collu

If the world is to be filled with innumerable discrete human level intelligences, the most plausible reason I can imagine them all secretly and flawlessly colluding to achieve the goal of destroying humanity (as opposed to poetry competitions or arguing amongst themselves like normal intelligent beings or selling ads like they were designed to do) is because their training data set is full of millenialist prophecy about AIs working together to achieve the [pretty abstract, non-obvious, detrimental in many ways] goal of destroying humanity from "AI Safety" fundraising pitches....