Human access to information has transformed a number of times. Before the internet physical publications were expensive to create. This ensured that credible and economically useful information was prioritized for publication. Books, magazines, and newspapers were few enough that the community of buyers and readers could feasibly police them for credibility. In the early years of the internet, this model was assumed to continue – finding information became harder and therefore search was popularized as an access channel, yet credibility was an assumed non-concern. With the explosion of online information, the reduction of publication costs to zero, and most importantly, the dramatic downscaling of user relationships with publishers, the community can no longer police the massive volume of tiny bits of information for credibility.
To address the problem, I build a prototype of a new search engine that uses credibility scores from expert review, combined with traditional popularity scores, to produce credible results for search queries. Early results show several percent of Google search results are from sources that are not factually oriented. Prototype access available upon request.
https://github.com/jjersin/cSearch
Individual search result:
Result score:
Result details:
