AUTOMATA: Um Ambiente para Combate Automático de Fake News em Redes Sociais Virtuais

Abstract

The propagation of fake news on social media has been increasing significantly in the last years. Despite the existence of applications that aim at suppressing the proliferation of fake news written in Portuguese, it was noticed the proposed solutions present a passive behavior regarding two aspects: (i) they are limited to identifying potential fake news only from content introduced by their users, as well as (ii) the absence of any procedures to combat disinformation. Therefore, this article presents AUTOMATA, an automated tool for combating fake news written in Portuguese. AUTOMATA periodically monitors posts made on social media and relies on Artificial Intelligence to detect suspicious fake news. After the detection process, AUTOMATA adopts a two-pronged approach to autonomously mitigate the widespread of this content, either by emitting posts on social media for warning about potential fake news or by sending the detected content to be curated by fact-checking agencies. This article also reports an experience in partnership with Ministério da Saúde for the application of AUTOMATA in the context of the COVID-19 pandemic in Brazil.