Wikipedia and fact-checking

Wikipedia's volunteer editor community has the responsibility of fact-checking Wikipedia's content. Their aim is to curb the dissemination of misinformation and disinformation by the website.

Wikipedia is considered one of the major free open source websites, where millions can read, edit and post their views for free. Therefore Wikipedia takes the effort to provide its readers with well-verified sources. Fact-checking is an aspect of the broader reliability of Wikipedia.

Various academic studies about Wikipedia and the body of criticism of Wikipedia seek to describe the limits of Wikipedia's reliability, document who uses Wikipedia for fact-checking and how, and what consequences result from this use. Wikipedia articles can have poor quality in many ways including self-contradictions. Those poor articles require improvement.

Large platforms including YouTube and Facebook use Wikipedia's content to confirm the accuracy of the information in their own media collections.

Using Wikipedia for fact-checking
Wikipedia serves as a public resource for access to genuine information. For example, the COVID-19 pandemic was an important topic on which people relied on Wikipedia for genuine information. Seeking public trust is a major part of Wikipedia's publication philosophy. Various reader polls and studies have reported public trust in Wikipedia's process for quality control. In general, the public uses Wikipedia to counter fake news.

YouTube fact-checking
At the 2018 South by Southwest conference, YouTube CEO Susan Wojcicki made the announcement that YouTube was using Wikipedia to fact check videos which YouTube hosts. No one at YouTube had consulted anyone at Wikipedia about this development, and the news at the time was a surprise. The intent at the time was for YouTube to use Wikipedia as a counter to the spread of conspiracy theories. This is done by adding new information boxes under some YouTube videos, thereby, attracting conspiracy theorists.

Facebook fact-checking
Facebook uses Wikipedia in various ways. Following criticism of Facebook in the context of fake news around the 2016 United States presidential election, Facebook recognized that Wikipedia already had an established process for fact-checking. Facebook's subsequent strategy for countering fake news included using content from Wikipedia for fact-checking. In 2020, Facebook began to include information from Wikipedia's info boxes in its own general reference knowledge panels to provide objective information.

Fact-checking Wikipedia
Fact-checking is one aspect of the general editing process in Wikipedia. The volunteer community develops a process for reference and fact-checking through community groups such as WikiProject Reliability. Wikipedia has a reputation for cultivating a culture of fact-checking among its editors. Wikipedia's fact-checking process depends on the activity of its volunteer community of contributors, who numbered 200,000 as of 2018.

The development of fact-checking practices is ongoing in the Wikipedia editing community. One development that took years was the 2017 community decision to declare a particular news source, Daily Mail, as generally unreliable as a citation for verifying claims. Through strict guidelines on verifiability, Wikipedia has been combating misinformation. According to Wikipedia guidelines, all articles on Wikipedia's "mainspace" must be verifiable.

Self-contradiction articles
A self-contradiction article is an article that contradicts itself.

An experiment was conducted on detecting self-contradiction articles on Wikipedia using a developed model called "Pairwise Contradiction Neural Network" (PCNN).

Contributions to this experiment are as follows:


 * A novel Wikipedia dataset named WikiContradiction was created which is the first dataset for self-contradiction tasks on Wikipedia.
 * A novel model PCNN was developed and was fine-tuned via the WikiContradiction dataset.
 * The empirical results exhibit the PCNN model's promising performance as well as highlight the most contradicted pairs.
 * The compiled WikiContradiction dataset can be used as a training resource for improving Wikipedia's articles.
 * This can further contribute to fact-checking and claim verification as well.

Limitations
When Wikipedia experiences vandalism, platforms that reuse Wikipedia's content may republish that vandalized content. In 2016, journalists described how vandalism in Wikipedia undermines its use as a credible source.

Vandalism is prohibited by Wikipedia. The website suggests these steps for inexperienced beginners to handle vandalism: access, revert, warn, watch, and finally report.

In 2018, Facebook and YouTube were major users of Wikipedia for its fact-checking functions, but those commercial platforms were not contributing to Wikipedia's free nonprofit operations in any way.

Self-contradiction limitations: The two main limitations of the self-contradiction PCNN model are the subjectivity of self-contradiction and not being able to deal with lengthy documents.