Blurring Political Leanings of Messages on Social Networks Using Deniable Steganography

Ming-Hung Wang,Po-Wen Chi

doi:10.1109/access.2019.2923975

Ming-Hung Wang, Po-Wen Chi

Open Access

https://doi.org/10.1109/access.2019.2923975

Copy DOI

Abstract

While social network services have become popular platforms where large amounts of user-generated content are published, individuals, election campaigns, and political organizations leverage the services to publish, promote, and exchange political messages. However, sometimes, political discussions resulted in heated debates, doxing, and even cyberbullying. These conflicts may decrease the engagement of users and prevent them from publishing articles with apparent political orientation. To help users to post political opinion among the Internet freely, in this paper, we employ the concept of deniable encryption to generate neutral-like articles composed of two critics toward both sides of the discussion, while one of the critics is the actual message from the author to particular recipients. For receivers who know the credential information (e.g., key and codebook), they can communicate with each other by extracting the real information from the neutral-like articles. For people with the opposite stand, they cannot claim the message that is supporting specific political entity, as it contains disapproval to both stands; besides, they cannot prove that which part of the message is the plaintext even if they intercept the messages. We provide a generalized process of our proposal and present a practical scenario using transcripts, Facebook posts of Donald Trump, and posts of Hillary Clinton to demonstrate the usability of our scheme.

Highlights

Social media and network sites have been used for individuals, the government, and political organizations to convey advocacy, promote election candidates, and monitor the public opinion
In this study, we propose a countermeasure toward such aggressive users using deniable encryption and text steganography concepts
Motivated by the abovementioned kinds of media content, in this study, we propose to leverage the above message forms to integrate the deniable concept on such textual content on social media sites

Summary

INTRODUCTION

Social media and network sites have been used for individuals, the government, and political organizations to convey advocacy, promote election candidates, and monitor the public opinion. We generate an article which is not easy for people to distinguish its political leaning, i.e., a counterfeit neutrality article, and containing the desired message in it The recipients can search the corresponding binary number of the keyword, indicating which part of the content is the actual message, to retrieve the message These deniable designs enable the author to deny the article has particular political stands, as it contains two critics. Though some neutral-like articles are generated through our design, all recipients in the same user group (owning the codebook) can correctly decrypt the stegotext, retrieve the actual message, and acknowledge the author of the article.

RELATED WORKS

FORMAL DEFINITION

IMPLEMENTATION

PRACTICAL SCENARIO

DISCUSSIONS

CONCLUSIONS AND FUTURE WORKS