ID: 2107
Presenting Author: DANIEL MAKALA
Session: 581 - Experiences in enhancing communication through the application of AI in IA
Status: pending
This study uses BERT and LSTM models to detect misinformation
on Chinese social media and assess its impact on the public
sentiment and economic perception.
Misinformation has become a serious social problem in recent years due to the rapid
proliferation of social media platforms, which allow large numbers of users to share
a wide range of content, including text, photos, videos, and audio. Misinformation refers
to false or misleading content that is broadly distributed. It can take various forms, including
fabricated stories, manipulated images, altered videos, and doctored audio, and can be
classified as disinformation, misinformation, or precisely. In China, platforms such as Weibo,
WeChat, and TikTok have become central to public discourse, yet they are vulnerable to rapid
spread of false information, particularly concerning social, economic, and political topics.
This study proposes a framework that combines deep learning models specifically BERT
and LSTM—for misinformation detection with a quantitative impact assessment on public
sentiment and economic perception. Using a dataset of 1 million social media posts collected
from major Chinese platforms, the study applies natural language processing, sentiment
analysis, and network analysis to evaluate the communicative influence of misinformation.
Expected outcomes include high-accuracy detection of misinformation, insight into sentiment
and engagement patterns, and a novel framework for linking detection to societal impact.
This research contributes to both artificial intelligence and communication studies, providing
a data-driven methodology for policymakers, regulators, and social media platforms to
mitigate misinformation effects in China and globally.
Daniel Makala is a researcher specializing in deep learning
applications in economics, and currently engaged in research
focusing on misinformation detection and impact assessment in
socia
Coauthor 1: DORIS KIWELU