[ThomPete]

This is pure speculation and not really very well thought through IMO.

We don't kill all ants either.

Even humans work with the principle of collateral damage with each other so if the idea is that AI would suddenly be without mistakes then that's already missing the point of what intelligence is. But just like humans any AI that is "super intelligent" will also need to have morals and ethics (or it can't figure out what IT want and doesn't want) and there is nothing that indicates they shouldn't be extensions of ours as we will be interfacing with them long before any superintelligence comes along, in fact it just as likely that it's our morals and ethics and culture that gets passed on and that biological humans don't get obliterated by AI but simply die out over time.

There is nothing bad about that, our genes survived.