Bias in Wikipedia
By Christoph Hube. 2017
50 citations recorded by the source corpus.
While studies have shown that Wikipedia articles exhibit quality that is comparable to conventional encyclopedias, research still proves that Wikipedia, overall, is prone to many different types of Neutral Point of View (NPOV) violations that are explicitly or implicitly caused by bias from its editors. Related work focuses on political, cultural and gender bias. We are developing an approach for detecting both explicit and implicit bias in Wikipedia articles and observing its evolution over time. Our approach is based on different factors of bias, with the most important ones being language style, editors, and citations. In this paper we present the approach, methodology and a first analysis.
Subfields: Communication, Information Systems, AI Ethics, Fairness & Society
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Communication
In Information Systems
In AI Ethics, Fairness & Society
Citations in the corpus
- Teaching language models to support answers with verified quotes (2022, background): “But it is not feasible to simply implement a trusted allowlist over sources: even relatively high-quality corpora like Wikipedia can contain errors or be biased (Hube, 2017; Martin, 2018), and no curated allowlist would cover all claims of interest.”
- What’s in the Box? An Analysis of Undesirable Content in the Common Crawl Corpus (2021, background): “Wikipedia (Hube, 2017), they have only scratched the surface of what these bodies of text contain.”
- Open-Source Embedding Models: A Comprehensive Survey of Techniques, Benchmarks, and Applications (2026, unknown): “Bias can also manifest in the portrayal of certain topics across different language versions, such as the negative emphasis on transgender individuals in Russian Wikipedia compared to English or French versions [55].”