User-Centered Detection of Fake News and Misinformation - Design and Prototypical Implementation in the System Contexter

Abstract

Misinformation or fake news may threaten our democracies, societies, and economies, even individual health and well-being. Humans are usually careful about the things they are being told. They check news or tweets against their knowledge or beliefs and estimate to what extent propositions contain information that is bogus. People have abstract representations of facts in mind. That help them to validate propositions and to search for information suitable for their validation. This paper presents design and prototypical implementation of the Contexter system that enables users to define and manage blueprints of facts or fake news. Contexter takes these blueprints as a schema to detect facts or fake news. It also starts to find variants of these blueprints to detect pieces of text that come semantically close to the propositions addressed by the original blueprint.