There is no "super intelligence", there is not even "artificial intelligence". These are just marketing terms.
In reality, all of AI ist just a bunch of programs, written by humans, and executed on a computer.
Thus all risks caused by AI are the result of *human* activity.
We need to educate the public about the nature of AI, now.
The public discourse is totally misguided and shaped by unrealistic fearmongering and other narratives that are out of touch with the technical reality.
The obvious sign of this is the talk about "dangerous" models.