I find the discussion about AI starting from the (mostly) for profit companies involved in AI very dubious. There are a many fields (bio-tech, nuclear, chemical) which could provide an "existential risks" to the human species, but most of the companies involved argue they control it, while other people argue they don't do it well enough.

I do think AI is posing a threat, but not "by itself" but mostly what it can do in the hands of misguided (or evil) people. The same as with previous risks.

And if our digital infrastructure can be brought down by something (AI/human/government) it is just badly done, being afraid of "AI" will not improve the situation. Someone should come with a proposal about what do we do now, but I would prefer more root cause solutions (how to make resilient systems) rather than "fear of AI". But building is hard and fear catches media attention, so...