In principle "who owns this jpeg" is just a few bits in a distributed filesystem that the community collectively agrees to treat as the canonical source of "ownership", and they could easily roll it back if someone stole a market-distorting amount of art.
In practice, if you do an interesting heist -- like you put on cool looking art thief costume and livestream yourself on a vintage PowerBook bypassing the owners' defenses and nabbing the apes with a narrow escape -- you've written a compelling story that the community is sort of bound to accept.
I disagree, since something that actually happened to you is anecdotal experience and therefore of very limited “good lesson” value.
An AI generated story that reflects and illustrates a data driven majority of experiences and resulting “lessons” would be much more valuable to me than your solitary true story, which may be a total statistical outlier, and therefore should not inform my decision making.
Kahneman explains it much better than I can, and in his book “Thinking fast and thinking slow”, he quotes studies and statistical analysis, how we as humans are commonly led to faulty decision making, because personal experience (“true stories”) tends to become our primary decision influencer - even if we have access to statistics that suggest the opposite of our own experience is the much more common experience.
So if the AI gives me access to a summarized better overall data based truth, wrapped into a made-up story (to help me remember better), then I would much prefer the AI to guide my decision making.