You might just need a large amount of click stream data from wikipedia, there is probably a cluster of people who trend to have such an awareness of the human condition. So it wouldn't require AGI to figure out a list of topics you should probably be aware of based on some sample topics. Although afaik such data is not stored by Wikimedia for privacy reasons, only less useful referrer data[1].

Another signal source might be based on the assumption that the important bits would be subject of suppression efforts, like topics that disappeared from school curriculum, internet censorship filter rules (for example [2]), takedown notices, banned books, topics journalists wrote about that were fired/defenestrated, wikipedia articles that were removed/renamed or heavily edited, unusual court cases with abnormal patterns might be recoverable from public records (like the Steven Donziger case [3]), websites on the internet archive that were taken down by infrastructure companies, there might be topics that are ignored by modern historians etc.

[1] https://meta.wikimedia.org/wiki/Research:Wikipedia_clickstre... [2] https://lumendatabase.org/faceted_search?sender_name=Europea... [3] https://en.wikipedia.org/wiki/Steven_Donziger