Starting to understand this AI thing now.
Basically, every benefit it will create has a negative that could be used by bad actors.So, cure a dangerous virus or make a dangerous virus, improve a countries defense strategy or create an attack strategy, make online banking more secure or delete it all with the backups, heal people or injure people, etc, etc.
So for every benefit you think about, or are fed, think of the negative reverse, more plentiful food supplies or starvation level scarcity. For every plus, there will be the same capability for an equally strong minus. For every step made towards a better future, AI will have the potential to enable the seeds of a worse future to be sown as well.
Super intelligence will be able to be used for good and bad in equal quantities, mostly invisibly. Not many of us understand much about our basic computers now!
The good is welcome but how does the world globally regulate the bad, which will inevitably end up in the control of a very few?
I should add, this isn't like managing the nuclear threat problem we have now. Only a few countries have nuclear arms. All countries, by its very nature, will have equal access to AI.