Spelling suggestions: "subject:"cocial both"" "subject:"cocial boys""
1 |
Twitter and social bots : an analysis of the 2021 Canadian electionDesrosiers-Brisebois, Gabrielle 12 1900 (has links)
Les médias sociaux sont désormais des outils de communication incontournables, notamment lors de campagnes électorales. La prévalence de l’utilisation de plateformes de communication en ligne suscite néanmoins des inquiétudes au sein des démocraties occidentales quant aux risques de manipulation des électeurs, notamment par le biais de robots sociaux. Les robots sociaux sont des comptes automatisés qui peuvent être utilisés pour produire ou amplifier le contenu en ligne tout en se faisant passer pour de réels utilisateurs. Certaines études, principalement axées sur le cas des États-Unis, ont analysé la propagation de contenus de désinformation par les robots sociaux en période électorale, alors que d’autres ont également examiné le rôle de l’affiliation partisane sur les comportements et les tactiques favorisées par les robots sociaux. Toutefois, la question à savoir si l'orientation partisane des robots sociaux a un impact sur la quantité de désinformation politique qu’ils propagent demeure sans réponse. Par conséquent, l’objectif principal de ce travail de recherche est de déterminer si des différences partisanes peuvent être observées dans (i) le nombre de robots sociaux actifs pendant la campagne électorale canadienne de 2021, (ii) leurs interactions avec les comptes réels, et (iii) la quantité de contenu de désinformation qu’ils ont propagé. Afin d’atteindre cet objectif de recherche, ce mémoire de maîtrise s’appuie sur un ensemble de données Twitter de plus de 11,3 millions de tweets en anglais provenant d’environ 1,1 million d'utilisateurs distincts, ainsi que sur divers modèles pour distinguer les comptes de robots sociaux des comptes humains, déterminer l’orientation partisane des utilisateurs et détecter le contenu de désinformation politique véhiculé. Les résultats de ces méthodes distinctes indiquent des différences limitées dans le comportement des robots sociaux lors des dernières élections fédérales. Il a tout de même été possible d'observer que les robots sociaux de tendance conservatrice étaient plus nombreux que leurs homologues de tendance libérale, mais que les robots sociaux d’orientation libérale étaient ceux qui ont interagi le plus avec les comptes authentiques par le biais de retweets et de réponses directes, et qui ont propagé le plus de contenu de désinformation. / Social media have now become essential communication tools, including within the context of electoral campaigns. However, the prevalence of online communication platforms has raised concerns in Western democracies about the risks of voter manipulation, particularly through social bot accounts. Social bots are automated computer algorithms which can be used to produce or amplify online content while posing as authentic users. Some studies, mostly focused on the case of the United States, analyzed the propagation of disinformation content by social bots during electoral periods, while others have also examined the role of partisanship on social bots’ behaviors and activities. However, the question of whether social bots’ partisan-leaning impacts the amount of political disinformation content they generate online remains unanswered. Therefore, the main goal of this study is to determine whether partisan differences could be observed in (i) the number of active social bots during the 2021 Canadian election campaign, (ii) their interactions with humans, and (iii) the amount of disinformation content they propagated. In order to reach this research objective, this master’s thesis relies on an original Twitter dataset of more than 11.3 million English tweets from roughly 1.1 million distinct users, as well as diverse models to distinguish between social bot and human accounts, determine the partisan-leaning of users, and detect political disinformation content. Based on these distinct methods, the results indicate limited differences in the behavior of social bots in the 2021 federal election. It was however possible to observe that conservative-leaning social bots were more numerous than their liberal-leaning counterparts, but liberal-leaning accounts were those who interacted more with authentic accounts through retweets and replies and shared the most disinformation content.
|
2 |
Act like a human, think like a bot : A study on the capabilities required to implement a social bot on a social media platform / Agera som en människa, tänk som en bott : En studie över de färdigheter som krävs för att implementera en social bot på en social medieplattformSamanci, Håkan, Thulin, Magnus January 2022 (has links)
Social media platforms have revolutionized how people interact with each other and how people gain information. However, social media platforms such as Twitter quickly became a platform for public manipulation and spreading or amplifying political or ideological misinformation. Although malicious content can be shared by individuals, today millions of coordinated automated clients disguised as individuals exist, also called social bots which have become a significant contributor of the malicious content spread on social media platforms. Therefore, this study aims to investigate in closer look what the requirements are to create a basic social bot from resources available from the internet. A proof-ofconcept prototype is implemented in the form of a social bot on Twitter. The experiences from the work indicate thatthe skills required to create a basic social botare well within reach ofa third-year Bachelor student in the field of computer science. / Sociala medier har revolutionerat människors sätt att integrera med varandra samt hur de delar och hämtar information. Däremot har sociala medier såsom Twitter snabbt blivit en plattform för offentlig manipulation och spridning eller förstärkning av politisk eller ideologisk desinformation. Även om skadligt innehåll kan delas av individer finns det idag miljontals koordinerade automatiserade klienter förklädda som individer, även kallade sociala bottar som har blivit en betydande bidragsgivare till det skadliga innehållet som sprids på sociala medieplattformar. Därför syftar detta arbete på att närmare undersöka vad som krävs för att skapa en grundläggande social bot på en social mediaplattform från resurser tillgängliga från internet. En bevis-på-koncept-prototyp implementeras i form av en social bot på Twitter. Erfarenheterna från arbetet tyder på att de färdigheter som krävs för att skapa en grundläggande social bot ligger väl inom räckhåll för en tredjeårs kandidatstudent inom området datavetenskap.
|
3 |
Twitter Bots as a Threat to Democracy : How political bots on Twitter jeopardized democratic functions of the online public sphere during the 2022 Swedish general electionWahlberg, Linus January 2022 (has links)
With more political and social discourse taking place online, particularly on social media, theorists have started labeling digital communicative realms as “online public spheres.” However, with the modern public sphere comes modern challenges to political communication; a core antagonist of which is political bots. Political bots are automated accounts that produce content and interact with individuals on political topics on social networks. In this thesis, I analyzed the presence of political bots on Twitter during the 2022 Swedish general election, and by examining the content posted by the bots, I investigated whether they jeopardized democratic functions of the online public sphere by publishing misrepresentation (i.e., artificially increasing the popularity of political actors and political ideas). The analysis uncovered significant bot presence during the 2022 Swedish general election; more than one-fifth of all election-related content was produced by bots, ~90% of which produced misrepresentation. I concluded that political bots jeopardized democratic functions of the online public sphere during the 2022 Swedish general election.
|
Page generated in 0.0557 seconds