Artificial Intelligence as a Tool for Basic News Literacy

Kiria, Ilya Vadimovich
PhD in Media Studies, Cand. Sci. in Journalism, Research Fellow, GRESEC Lab, University Grenoble Alpes, Grenoble, France, Professor, Institute of Media, HSE University, Moscow, Russia
Professor Institute of Media, National Research University Higher School of Economics
ikiria@hse.ru

Shomova, Svetlana Andreevna
D.Sc. in Political Sciences, Professor, Institute of Media, HSE University
sshomova@hse.ru

Ilvovsky, Dmitry Alexeevich
Cand. Sci. in Technics, Assistant Professor, School of Data Analysis and Artificial Intelligence, Research Fellow, International Laboratory for Intelligent Systems and Structural Analysis, HSE University
dilvovsky@hse.ru

Dushakova, Irina Sergeevna
Cand. Sci. in Philology, Senior Lecturer, Institute of Media, HSE University
idushakova@hse.ru

Abstract
This talk addresses the results of the development of a computer tool that allows to evaluate to what extent a particular news text contains distorted information and is a biased message about a particular event based on text markers. Thus, we present a tool that is still tested that allows average people to check how reliable a particular message of the mass media is. Under current conditions of the “information littered” space, the possibility to draw this distinction is an extremely important competence.
The talk describes the preparation of a dataset for such an instrument that works with the Russian language. This set contains about 100 texts taken from the Yandex.Zen resource. The texts were annotated: fragments that contain distorted methods of presenting information (in English terminology—computational propaganda) are highlighted. A list of the most commonly used techniques was created for this research; it is based on similar typologies for other languages. Fragment detection and technique classification experiments were performed on the new dataset using models that provided best results on similar datasets for English.
The report also presents the first results of creating a tool that allows checking the personalities of experts referred to in news reports.

Keywords: fake news, computer propaganda, media bias, news literacy