I’d like to respectfully disagree that humans need an agency of any sorts, and I say this because I am a firm supporter of replicating nature in human systems (biomimicry). As far as I know, animals and plant life don’t rely on agencies to interact with each other. Humans have been hoodwinked into believing that we can’t be accountable for our own social engagements and that we need agencies (middlemen) to resolve our issues.
This is the main reason I am so excited about blockchain — it allows us to remove the middlemen (agencies) and, through the magic of smart contracts, imagineer a society without any form of agency. Admittedly we have a long way to go, but why not start the conversation now?
As for the characteristics of a future you have asked about, we are planning an event for February 2020 along the lines you have suggested, except that AI models won’t feature as strongly as you suggest (it’s too early and AI isn’t well-developed enough yet). What your AI vision is replaced with is a library of simulations so that voters can model what the outcomes of their votes will likely look like, tapping into Big Data. The event aims to answer the question “What does it mean to be human in the 21st century?” The outcome is intended to be a new social contract for humanity, which we then track the implementation of over the next 10 years (publicly). The entire process will be livestreamed and open source. And global. We are following the model Iceland followed in 2011.
This is not public yet, but if you’re interested, I’l be happy to send you a link to the planning documents privately — just let me know.