Artificial Intelligence in Fake News Detection and Analysis for Low-Resource Languages

The dissemination of fake news and the development of public faith in it are both influenced by psychological and sociological factors. After repeated exposure, various people start trusting this fake news. So, there is a need for in-depth analysis by having a perception of who creates fake news, how and who created it, along with its way of propagation and the motivation behind detecting it. This also creates a need for an automatic fake news detection system that is independent of language and can be low resource or high resource. This paper includes the general framework for a fake news detection system that gives an idea of the various steps involved in it, each of which has its own significance. It is found that the main source of spreading fake news is social media, but other sources are also available for the purpose of testing the proposed system on data collected from different sources, and each type of data has different features that can be understood from the study done in this work. In automatic detection, an important role is played by artificial intelligence and machine learning on various languages, such as Arabic, Italian, Dutch, French, Japanese, Korean, Slavok, Polish, Spanish, German, Swedish, English, Urdu, Dravidian, Kurdish, and Turkish covered in this work. Furthermore, a comparative analysis is done on it in terms of different datasets, purposes of work, extraction or classification approaches used, and results obtained using them that help other researchers get the best results in their work.