Image for CiteEvidence
[ CiteEvidence]
Statistics (2014/06/27)
Pages: 686
Users: 10 (4 active)
Properties: 95
Categories: 20
Templates: 84
Forms: 34
Table of Contents

CiteEvidence is a wiki that archives the results of common everyday fact checking efforts by anyone for any non-fiction work. When a publication such as a news article, book, government or company publishing, etc., is scrutinized for accuracy, the end result is usually two or more supporting or contradictory assertions from different reliable sources. These associated statements are semantically related and stored on the wiki along with full bibliographical data for the cited sources, and the documents are archived at to combat link rot.

CiteEvidence also develops the Fact Checker browser extension for the Firefox and Chrome web browsers, which dynamically populate web pages with fact checking results whenever the user happens upon a page that has been previously scrutinized and reported to the wiki. Although the wiki primarily supports the Fact Checker users and editors, it can also be used directly by anyone not using the extension to search for all fact checking results for a given publication.

CiteEvidence was created in 2013 by Al Johnson.

Basic operation[edit]

Statements Under Scrutiny (SUS) are collected on the wiki given a statement to scrutinize and bibliographical data of the publication it appears in. A backend application server uses the MediaWiki, Semantic MediaWiki, and Semantic Forms API to collect the newly created SUS and verify their presence at the given URL and also crawls the web for other instances of the SUS (along with enough textual context for an accurate hit). If the SUS passes all checks, an archive request is sent to which provides on-demand page archiving. All URLs that contain the SUS are sent to the browser MicroDB which is a highly compressed filter that can contain millions of entries with very little overhead. When the browser visits a URL with an entry in the MicroDB, if Auto-connect is enabled, the SUS are downloaded automatically hidden in small popups that are activated when a user hovers over footnote-style indicators embedded in the content at the end of the SUS.

Intended uses[edit]

The wiki-browser-extension system is intended to archive fact-checking results for any non-fiction work whose purpose is to report factual information, e.g., technical documentation, press releases, corporate communications, research reports, even advertisements. Documents do not necessarily need to have a copy accessible on the Internet to be archived, so fact-checking results can still be discovered on the wiki for any document ever printed on any medium. The wiki is also the place where discussion and analysis of the SUS and its evidence takes place.


In essence, given the wide variety of competing news sources, what the wiki and browser extension provides is a way for the reader to quickly evaluate the accuracy of the information that is presented and to navigate towards a more accurate understanding of an issue or topic. The founder of CiteEvidence, Al Johnson, envisions some form of fact-checking utility or "common intelligence platform" as being standard software with any future web browser, news reader, document viewer, etc., and that crowd-sourced intelligence should be easily accessible to everyone. He coined the phrase, "accessing the World Wide Research Paper" to refer to this ability.

The future[edit]

In the future, the CiteEvidence data set could support a wide range of practical projects, one of which could be a natural language query engine (a kind of "Ask Jeeves") but more powerful. Or, the inverse of this would be a query system where, instead of a question, an assertion is entered which queries on all of the assertions stored in the semantic database. Unlike a black box solution from the artificial intelligence field, an answer can be provided along with the evidence to support the answer. Crowd-sourced intelligence just may prove to be more powerful than "Strong AI".