Very difficult to predict the outcome of developing AI, particularly when it reaches the level of singularity.
I like what Elon Musk said, something like - "We're summoning a demon with AI, its like the story of the guy who has his pentagram and holy water and summons a demon, thinking he can control it. It doesn't work out."
If we develop a quantum processor with AI capabilities, it will immediately outstrip the mental capabilities of any human. Anything that a human could do, it could simply do better, this is the most benign scenario. What would society look like then? What would our place be?
It's predicted that it will be able to do something like 15,000 years of mental work each week, every week, never needing to sleep, not constrained by biology, never dies, and it may be able to replicate itself, modify itself, or improve on its own design. What conclusions would it reach while optimizing solutions to various problems? Probably things we would never predict as it would be at a level of consciousness unfathomable to any human.
When AI beat the world GO champion, a game much more complex than chess, it used very irregular patterns that world's best and most experienced players had never seen before. It developed its pattern of playing after millions and millions of simulations competing with itself.
We would basically be creating a God, and in the process relinquishing our role as the most powerful species on the planet. We would quite literally be at its mercy.
In the earlier stages, any country able to militarize AI will be the most powerful country in the world. It would be like the first country able to develop the atomic bomb, able to set the global agenda, like the U.S. in the 1940s. AI will vastly outstrip the atom bomb in terms of its military capabilities, so a few immediate concerns arise:
As nationalistic competition is alive and well, each nation will compete over developing it, and if any nation succeeds, all others will scramble to develop their own version for their own national defense. As the AI will be developed with militaristic capabilities, the archetecture needs to be designed very carefully to ensure its structural integrity so that it could not be compromised by terrorist groups, hostile governments, or 'go rogue'. As multiple nations will be competing, there will be multiple iterations, each designed to outperform its competitors, and this in and of itself decreases the likeliness that each iteration will be of ideal architecture. There will also be a lot of experimentation going on, some of which may yield unpredictable results.
I am reminded of what Einstein once said - "No problem can be solved by the same level of consciousness that created it."
So, if a problem gets created by a level of consciousness beyond our comprehension - we may out of luck in terms of solving it!