Avaliação de comunicabilidade da ferramenta de detecção automática de fake news, FakeCheck, em plataforma WEB

Abstract

The spread of fake news has become a growing problem in society, directly impacting public opinion and influencing political and social events. To mitigate this phenomenon, several artificial intelligence (AI)-based solutions have been developed, such as FakeCheck, a tool that uses AI techniques to automatically classify news as true or false, with 90% accuracy. However, beyond algorithmic precision, the effectiveness of these solutions also depends on how their interface communicates with users, ensuring they can navigate them easily and understand the presented results. This study evaluated the communicability of FakeCheck on the web platform using two methods from Semiotic Engineering: the Semiotic Inspection Method (SIM) and the Communicability Evaluation Method (CEM). SIM analyzed the interface from the designer’s perspective, while CEM observed real user interactions. The results indicated that, although the tool is intuitive, there are communication flaws in the interface, such as the lack of explanations about detection models and difficulties in interpreting feedback. Furthermore, the analysis highlighted the need to refine messages and include new elements to enhance communication between the designer and the user, reducing communicability breakdowns. The comparison between methods allowed for identifying both design flaws and user interaction challenges. This study is expected to contribute to the improvement of FakeCheck and similar tools, making them more accessible and functional.