Your post covers various areas but some counter points:
 
Job loss via robotics does not fall under job loss via AI. They are currently separate categories that in the future will have more overlap.
 
Automation has previously directly referenced automation via robotics. Automation term will be expanded to also include AI but this is quite new and most people do not use automation to cover both until very recently..
 
I wouldn't worry about government adoption of AI. All it takes is 1 locality / city / state or nation to adopt AI and yield strong results that outpace previous metrics which will create the incentive to others to adopt.
 
Economic systems will no longer be defined by capitalist / communist / socialist / etc. A new economic system will arise through maximizing equality / output and incentive for the individual. The 3 pillars of measuring an economic system. I suspect an AI to have the ability to increase all 3 over previous systems and the economic system will not fit a human economic system of the past.
 
Adoption of AI will happen regardless as its too useful of a tool. The problem is who gets to code the AI which determine its output at this point and also the alignment problem if it becomes sentient which at this point is nothing more than a dream rather than something routed in reality.
 
Job losses can be lowered by many roles being augmented by AI to increase output in the short term. However eventually job losses will occur forcing the adoption of UBI. Humans will be retained to do jobs the AI can't which will become smaller and smaller gaps over time. More and more people will end up on UBI. No idea what happens after that.
 
AI will be taking smaller roles first in society, showing that its better than humans in those roles, humans will expand the role of AI to cover larger responsibilities like governance / the application of law and economy. How else are people going to have faith in AI?
 
I am not concerned about the creation of sentient AI. As I have no power to stop or slow its progress. I believe the creation of sentient AI is a requirement for a better future. However I also think its a lot harder than the people in this subreddit believe and will take longer. I have a lot of faith AI's we build between a sentient AI and now will greatly help humanity in the meantime.