I'd think the most relevant academic language would come from clinical psychology, since it boils down to a fear of hypothetical and somewhat-unreasonable vengeance.

No technology needed: It's the same as a prophesized king, believing his own divinity, who orders his knights chop-up everyone who didn't actively try to help his parents marry and fulfill the prophecy.

Roko's Basilisk is a modern version of Pascal's Wager. The same counterarguments works. Just search-replace God to AI, are you are done.

That's an interesting question. I don't know the answer.

It reminds me of a video by Yuval Noah Harari https://youtu.be/1rtS2OEV6bM?si=WNuCtsr92O-06jyQ where he argues that, from an ethical and political point of view, the basis of consciousness is the potential for suffering. Only the conscious can truly suffer.

Only the conscious can be punished. So, answering your question implies defining non-human agents that are conscious. That's not just non-human technological agents, animals qualify too as non-human agents.