Fighting misinformation in the digital age: a comprehensive strategy for characterizing, identifying, and mitigating misinformation on online social media platforms
Date
2023-09-25
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Veebipõhiste sotsiaalmeediaplatvormide, nagu Twitter ja Facebook, esilekerkimine on hõlbustanud valeteabe ülemaailmset levitamist, soodustades sotsiaalse hirmu, ärevuse ja majandusliku kahju kasvu. Lõputöö uurib mitmekülgset lähenemisviisi desinformatsiooniga võitlemiseks digiajastul, keskendudes kolmele põhidimensioonile: valeinformatsiooni sisu tuvastamine, raamistiku väljatöötamine valeinformatsiooni levitajate tuvastamiseks, ja tõhusate desinformatsioonivastaste meetmete rakendamine.
Esiteks on meie väljapakutud postituste iseloomustamise meetodi eesmärk mõista kuulujuttudest ja mittekuulujuttudest postituste tunnuseid, et tuvastada postitajate kognitiivne tegevus ja desinformatsiooni levitamise motiivid. Sotsiaalmeediapostituste omaduste põhjalik uurimine aitab teadlaskonnal tuvastada ja vältida desinformatsiooni.
Teiseks ei ole varasemad meetodid kahtlaste või pahatahtlike kasutajate ja desinformatsiooni tuvastamiseks Twitteris ja teistel sarnastel platvormidel piisavalt kaalunud kasutajatasandil toimuvat tuvastamist. Ühe postituse põhjal kasutaja kuulujuttude levitajaks liigitamisest ei piisa. Meie panus sellesse valdkonda on klassifitseerimisraamistik, mis ühendab parema lähenemisviisi väljatöötamiseks mitmed postitused ja võrguteabe.
Kolmandaks on olemasolevad sotsiaalmeedias desinformatsiooni leviku piiramise lähenemisviisid kohati piiratud, näiteks puudub väline modereerimine ja süsteem tugineb rangetele eeldustele. Esitame automatiseeritud lahenduse valeinformatsiooni suuremahuliseks ümberlükkamiseks, kasutades selleks sotsiaalmeedia andmeid ja kureeritud kontrollitud faktidega andmehoidlaid. Eelkõige keskendutakse selles aspektis Twitteri platvormile ja COVID-19 väärinfole, uurides kahte teineteist täiendavat lähenemisviisi.
The emergence of Online Social Media (OSM) platforms like Twitter and Facebook has facilitated the global dissemination of false information, contributing to social fear, anxiety, and economic harm. This thesis examines the multifaceted approach to combating misinformation in the digital era, focusing on three key dimensions: identifying misinformation content, developing a framework for identifying misinformation spreaders, and carrying out effective counter misinformation measures. First, our proposed characterization approach aims to understand the traits of rumor and non-rumor posts to identify cognitive behavior and motivations behind spreading misinformation. Scrutinizing social media posts' characteristics helps the research community detect and prevent misinformation. Second, previous techniques for detecting suspicious or malicious users and identifying misinformation on platforms like Twitter have not considered user-level detection adequately. Categorizing a user as a rumor based on one post is insufficient. Our contribution is a classification framework that combines multiple posts and network information to develop a better approach. Third, existing approaches for mitigating the spread of misinformation on social media have limitations, such as the lack of external moderation and reliance on strict assumptions. We propose an automated rebuttal of misinformation at scale by harnessing social media data and curated fact-checked data repositories. Particularly, this dimension focuses on the Twitter platform and COVID-19 misinformation, exploring two complementary approaches.
The emergence of Online Social Media (OSM) platforms like Twitter and Facebook has facilitated the global dissemination of false information, contributing to social fear, anxiety, and economic harm. This thesis examines the multifaceted approach to combating misinformation in the digital era, focusing on three key dimensions: identifying misinformation content, developing a framework for identifying misinformation spreaders, and carrying out effective counter misinformation measures. First, our proposed characterization approach aims to understand the traits of rumor and non-rumor posts to identify cognitive behavior and motivations behind spreading misinformation. Scrutinizing social media posts' characteristics helps the research community detect and prevent misinformation. Second, previous techniques for detecting suspicious or malicious users and identifying misinformation on platforms like Twitter have not considered user-level detection adequately. Categorizing a user as a rumor based on one post is insufficient. Our contribution is a classification framework that combines multiple posts and network information to develop a better approach. Third, existing approaches for mitigating the spread of misinformation on social media have limitations, such as the lack of external moderation and reliance on strict assumptions. We propose an automated rebuttal of misinformation at scale by harnessing social media data and curated fact-checked data repositories. Particularly, this dimension focuses on the Twitter platform and COVID-19 misinformation, exploring two complementary approaches.
Description
Keywords
false information, social media, social media platforms, information society, information technology, information processing, data processing