This is a brief description of an idea to solve the fake news and misinformation problem. Longer take and more details are in in design notes file.
Determining what’s “true” is hard, often important and should be easier. We do it unconsciously, not-objectively and under the influence of the media, authorities, our friends or foreign actors - our “social bubble”.
It was always bad, but with the current state of fake news and public discussion we need better solutions.
Learning what’s true is difficult on some topics, and learning WHY we believe it’s true, or how our understanding changed in time requires a lot of manual research.
Solutions vary and are not perfect: neutral Wikipedia with extensive rules, peer-reviewed science journals, social media Facebook/Twitter/Reddit - with a non-transparent moderation and “we only deliver text” approach or public media radio/tv under government influence).
There are personal knowledge graph applications, centered around a particular research topics of a single person. Conceptually similar, but unable to perform on a large scale and lacking simple cooperation or social functions. See for a reference: Org (org-roam), Logseq, Obsidian, Roam research, Zettelkasten.
System tracking atomic bits of information, sources, authors, votes and relations between them, in order to help evaluate the information (eg. 82% consensus that it’s true) in respect to your explicitly chosen and visible “bubble”.
You can trust your friends, or universities, or wikipedia or certain news sites. Your trusted sources (eg. uni) might trust some other sources (eg. a publication author) and by extension you trust them. Using those relations system can decide whether information is true or not.
You can get a simple answer, but can dig deeper to see “why”, and the bubble is visible and transparent. You can alter your trust, change opinion and see how it reflects the results.
Example: “current inflation levels will help economy by 2023” and you could see what your trusted people think about it (by voting, or trusting other parties), what government thinks and what the opposition believes. Or, say, evaluate whether “bleach/hydroxychloroquine cures covid-19” is true.
System tracks a graph of knowledge and a social-trust graphs which it can intersect to evaluate various metrics.
For instance, flatearthers and their statements could exist on this platform, but it would be explicitly visible how separated their beliefs are from the current state-of-the-art science. Sources and proofs of their statements could be tracked and compared to the sources and proofs for spherical earth.
- Even if moderation of the content itself could be deemed not necessary, it still would be necessary to control the spam and bots and artificial, automated fights over the content. Social trust, account authority could help it a bit maybe.
- A lot of content would have to be generated and tracked for the site to be operational and useful. That would be a wikipedia-level effort to get going.
- Data would have to be linked and deduplicated, people are not good at this.