Fake News Detection System using Web-Extension

Abstract

Internet is a supreme one-stop source of information that enables the sharing of news and curated user-content at a rapid, effortless, and in a routine manner. News is a global medium of daily events worldwide, offering absorption of quick information. With ample availability of news content online, these news articles has by-products in information generation in both ways -real and fake news. Considering the context and volume of information shared online, it is challenging to establish authenticity of news. This leads to the immense growth of fake news on various websites, which can lead to serious concerns in society, fading away the correct news content to reach the users creating misconceptions and deceived views of the readers. To ensure the readers have the credibility of the content, we propose a web-based extension enabling them to distinguish from the fake and real news content. The proposed web extension in the paper uses multiple deep learning models. The first is based on our model trained on LSTM, and the other uses OPEN AI’s well-developed AI-generated text classifier GPT-2. The devised web-extension displays both probabilities of news being either AI-generated or written by an individual.