Tackling Multilinguality and Internationality in Fake News

Proceedings of Recent Advances in Natural Language Processing, pages 1380–1386 Sep 1–3, 2021. https://doi.org/10.26615/978-954-452-072-4_154 1380 Tackling Multilinguality and Internationality in Fake News Andrey Tagarev Research Unit, Sirma AI 79 Nikola Gabrovski St Sofia, Bulgaria andrey.tagarev@ontotext.com Krasimira Bozhanova Semantic Analytics Solutions, Sirma AI 79 Nikola Gabrovski St Sofia, Bulgaria krasimira.bozhanova@ontotext.com Ivelina Nikolova-Koleva Bulgarian Academy of Sciences and Sirma AI 79 Nikola Gabrovski St Sofia, Bulgaria ivelina.nikolova@ontotext.com Ivan Ivanov Life Science Solutions, Sirma AI 79 Nikola Gabrovski St Sofia, Bulgaria ivan.b.ivanov@ontotext.com Abstract The last several years have seen a massive in- crease in the quantity and influence of disin- formation being spread online. Various ap- proaches have been developed to target the process at different stages from identifying sources to tracking distribution in social media to providing follow up debunks to people who have encountered the disinformation. One common conclusion in each of these ap- proaches is that disinformation is too nuanced and subjective a topic for fully automated so- lutions to work but the quantity of data to pro- cess and cross-reference is too high for hu- mans to handle unassisted. Ultimately, the problem calls for a hybrid approach of human experts with technological assistance. In this paper we will demonstrate the applica- tion of certain state-of-the-art NLP techniques in assisting expert debunkers and fact check- ers as well as the role of these NLP algorithms within a more holistic approach to analyzing and countering the spread of disinformation. We will present a multilingual corpus of dis- information and debunks which contains text, concept tags, images and videos as well as var- ious methods for searching and leveraging the content. 1 Introduction The topic of fake news and intentional spread of disinformation has been gaining increasing promi- nence over the last decade. The distinction of termi- nology and attempts to classify various erroneous or misleading statements online is constantly evolv- ing but disinformation, as shown in Fallis (2015), has recently settled as the commonly accepted term to describe the intentional and systematic spread of incorrect information. The spread of this incorrect information is strongly reliant on social media, causing a strong emotional reaction and quick propagation before its inaccuracy can be effectively exposed. This means that in most cases minimal effort is put into craft- ing the disinformation, instead relying on speed, volume and reuse of slightly modified pre-existing materials. There are, of course, always new disin- formation materials popping up but they are in the minority and should they gain traction, they will very quickly get picked up, modified slightly and reused. A very stark example of this kind of interaction was provided in the early months of the Covid pan- demic as reliable scientifically-tested information was still rather scarce and the void was filled by a wide variety of rapidly-spreading fake and unsup- ported claims. This can be viewed both from the perspective of journalists mobilizing to counteract the spread1 and from that of researchers looking into assisting their efforts e.g. in identifying spread of disinformation that has already …

Content truncated to 500 words