This is NOT a role for the OSM[0]. In many cases it is the best source of map data, but it will never be the sole single authoritative source of truth.
Treating it as one is horrible idea and will end in tears.
[0]or Google Maps
And often the things you want to map are better mapped seperately in their own dataset. Just consider woodland. Should that block of trees be mapped as individual trees, a block of woodland, or as a linear hedgerow? Different people would answer that question differently. A topographic map forces everyone to have the same answer.
I think that citizen mapping should move to focus on particular subjects and themes that the individual is interested in. And try and get closer to the underlying phenomena that is being described.
Improving OSM is an area of active research I believe, but I'd suggest the scale of the challenge is of a different magnitude than can be met by the type of ideas you mention. Think about something like Google Maps, or in the UK OS MasterMap. Each of these likely has hundreds of millions of R&D invested in tools, techniques, processes and infrastructure behind them, as well as an ongoing budget in the hundreds of millions, and numerous, well organised, skilled, full time staff. There's a critical mass behind such efforts that as yet OpenStreetMap has not been able to muster, which is no criticism of OSM, but just an acknowledgement of the reality of the situation. We're talking about setups that take dedicated organisations years to develop - the history of the Ordnance Survey for example goes back to 1747!