Answer the pretender, fifth of December
So having been to a talk by Audrey Tang yesterday (look her up if you never heard about her), I'd like to share some interesting points:
Human-in-the-loop does sound like it could make us end up in a hamster wheel of ever faster turning (or churning) AI. What makes a lot more sense is AI-in-the-human-loop (where the AI is used to facilitate the human processes).
Do not delegate what you cannot delegate. It make as little sense for a robot to go to a gym as it does for AI to replace civic processes among humans, like for example care or government processes.
Prevent digital colonialism and nationalism. They promote polarisation and imprison you. Build bridges (between platforms and people) so we don't imprison ourselves. If necessary, force companies to comply by throtteling their connection (it's not censorship if it just becomes slow).
Find the uncommon ground. People agree on way more than you'd think, so figuring out where they (unexpectedly) agree can help a lot to build solutions that accept the lines of disagreement and still work.
The singularity may be near, but the plurality is here.
(Audrey Tang if it wasn't already clear)