This is far outside the Overton window, but I think it will be what wins out in the end. All information deserves to be free.
The society that used it will grow with the value of the information. That value can then be captured through increased land rents. Those land rents can then be used to pay back the creator through either a lump sum or a residual process.
Since all information released in this process is free-to-use, new information will flourish just like the eco-system behind stable diffusion is flouring faster than OpenAI.
That's an implementation detail. Do we first agree that the system proposed would lead to a better society? Would lead to the pie growing by a lot? Would stop heaps of wasteful processes?
Since people already give away information for free - see open source software, wikipedia, all sorts of other sources, you wouldn't need to be even close to accurate to start with. Can we agree that Wikipedia provides at least $1b/year in value? Great, give them $100m/year and we'll work from there. Linux? Let's go with another $100m. Firefox? Stable Diffusion? That guy in Nebraska[0]?
I'm sure someone else can come up with a better implementation, but a way is through a very small subset of people, when trying to download a book or use other information, they get entered into an auction shared by some others who also want to use that information. Anyone who bids more than the median gets to use the information and pays the median, anyone who bids less than the median doesn't - but gets the median value in return.
The society that used it will grow with the value of the information. That value can then be captured through increased land rents. Those land rents can then be used to pay back the creator through either a lump sum or a residual process.
Since all information released in this process is free-to-use, new information will flourish just like the eco-system behind stable diffusion is flouring faster than OpenAI.