Depends on how much power we let these bastards have. In western democracys, power restriction is our most important mechanisms.
But feel free to ask ai these questions
But these power restriction measures in the west were not implemented as a choice based solely on ethics or rationality, but rather because societies become less efficient under dictatorial-type governments. During WW2 and after many less-democratically inclined governments got naturally selected by democratic nations as those are technologically and economically superior (democracies and or capitalism leads to internal competition, which leads to more technological advances, which leads to more military power, etc.). This happened naturally, and not because a select group of people out of the blue chose for it to be so.
Similarly, it is infeasible to place restrictions of the development and or capabilities of AI. AI has a lure of immense power; physically, financially and militarily. Any government/company/military will not want to stop these developments for themselves, as then another entity will go and get it and get the upper hand. Let alone that it will be introduced as something positive (AI will eliminate the need for people to work, it will cure cancer, etc. etc.), rather than framing it in the broader picture (a cog in the techno-industrial system which enslaves humanity and removes all individual autonomy), leading to AI developments having effectively no resistance, especially not by those who pull the strings.
So to say that we could place limits on AI is in my opinion kind of naive. It will take it's natural course, and given the complexity of our world it is impossible to say what that exactly entails, but it's not likely to be anything positive for the human race (or any of the life on earth).