I'm not the grandparent but the more interesting question is what could possibly constitute "supporting evidence" for an AI Doom scenario.

Depending on your viewpoint this could range from "a really compelling analogy" to "A live demonstration akin to the trinity nuclear test."