Deepfake em vídeo como estratégia de convencimento : a estética na era da desinformação

Abstract

Faced with the increasing access to the internet, the use of social media occupies space in the daily life of a significant portion of society. The environment is conducive to the dissemination of false information, called Fake News, whose technological advances have taken them to another level. The increasing sophistication of AI from mobile cameras and the high reach of social media have made the creation and propagation of edited videos even more compelling. These videos are called deepfake. In Post-Truth times, where concrete facts have less influence than preconceived emotions and beliefs, it is undeniable that regardless of their veracity, facts have the capacity to impact society. This work seeks to analyze the convincing power of synthesized videos, especially in the political sphere, to measure the potential risks imposed on society by fake videos and what this implies. Based on the analysis of aesthetic aspects of the best-known deepfakes in Brazil, this work proceeded to the application of forms and questionnaires, in which it was possible to observe that old videos, with less visual definition, can be potentially more reliable and convincing than current videos and of better quality. Finally, this work points to necessary reflections on the future of creating videos from Artificial Intelligence.