Desde la llegada del primer caso de COVID-19 en el Perú, diversos medios de comunicación han emitido información sobre la pandemia. Asimismo, durante este contexto, ha existido una serie de disputas entre el poder legislativo y ejecutivo. En ese marco, el 14 de noviembre murieron dos jóvenes en la protesta contra el expresidente Manuel Arturo Merino de Lama. Después de esta trágica sucesión, el programa “Rey con Barba”, del canal Willax TV, brindó cierta data sobre la protesta que fue cuestionada. Del mismo modo, el otro programa “Beto a Saber” proporcionó información sobre la vacuna contra la Covid-19 que fue discutida por diversos medios, ya que fueron catalogadas como fake news. Para poder evidenciar la existencia del componente de las fake news en ambos programas se realizó un decoupage; el cual cuenta con 3 categorías: categoría audiovisual, categoría discursiva y categoría de las fake news. / Since Covid-19 arrived to Perú, several networks have been displaying information about the pandemic. At the same time, there have been plenty discussions around the legislative and Judicial branches. On November the 14th, two young men died during the protest against the past president Manuel Arturo Merino de Lama. After this tragic event, the program “Rey con Barba” from the Willax TV channel, which shared controversial information about the protest. Additionally, the “Beto a Saber” program, the introduces information the Covid-19 vaccine which was debated between different medias, labeled as fake news. In order to provide evidence about the presence component of fake news in both programs, a decoupage has been done, witch consistent at three categories: the audiovisual category, the discursive category and the fake news category. / Tesis

El presente trabajo analiza los factores que influencian la viralización de las fake news en Twitter durante la pandemia por Covid-19. En la actualidad, por la pandemia generada por el Covid-19, las fake news han sido tema de viralización en redes sociales. Por un lado, el coronavirus al ser una enfermedad nueva, ha sido utilizada para la creación de noticias falsas, ya que aún no hay información concreta y veraz por lo que es fácil crear y especular rumores. Por otro lado, ha salido de nuevo a la vista las fake news, ya que el uso de la internet ha generado que este fenómeno se vuelva más viral y tome relevancia en la actualidad, además de que por la pandemia muchas personas han tenido que cambiar su vida “normal” por una “virtual”. Las Fake News es un término que se le da a las noticias falsas que provocan alarma y desinformación en las personas, añadiendo que son más propensas a iniciar en internet. Lo que esta investigación busca a través de un análisis de contenido es hallar los factores que influyen en estas noticias y el porqué de su viralización. Asimismo, se seleccionarán tweets de Twitter que utilicen el hashtag #YoNoMeVacuno para entender a los usuarios de esta plataforma y ver los efectos de la viralización de Fake news en esta red social. / The present work analyzes the factors that influence the viralization of fake news on Twitter during the Covid-19 pandemic. Currently, due to the pandemic generated by Covid-19, fake news has been the subject of viralization on social networks. On the one hand, the coronavirus, being a new disease, has been used to create false news, since there is still no concrete and truthful information, so it is easy to create and speculate rumors. On the other hand, the fake news has come out again, since the use of the internet has caused this phenomenon to become more viral and become relevant today, in addition to the fact that due to the pandemic many people have had to change their "normal" life for a "virtual" one. Fake News is a term given to this news that causes alarm and misinformation in people, adding that they are more likely to start on the internet. What this research seeks through a content analysis is to find the factors that influence this news and the reason for its viralization. Likewise, Twitter tweets that use the hashtag #YoNoMeVacuno will be selected to understand the users of this social network and see the effects caused by the viralization of Fake news on this platform. / Trabajo de investigación

The rise of social media platforms has had a significant impact on the conventional model of gatekeeping. With increased access to information--as well as the ability to contribute to the public discourse--individuals no longer need to rely on the mass media for news. These realities have led to increased conversations surrounding credibility in the digital age. Although not a new concept, fake news has become increasingly common in recent years. The web--particularly social media outlets, like Twitter--have enhanced the spread of misinformation. To combat this, social media platforms have introduced gatekeeping features like verification marks and warning labels. However, questions remain regarding the credibility and effectiveness of these features. Furthermore, little information exists regarding the perceptions of these features. For this study, the researcher examined the perceptions of fake news, verification, and Notices (i.e., warning labels) as they relate to Twitter. These perceptions were captured through a survey that was distributed to Twitter users through MTurk. Results were examined generally as well as in the light of political orientation, ranging from very liberal to very conservative on a 4-point scale. Within the scope and limitations of this study, results indicate that the majority of Twitter users believe that fake news on the platform is a major problem. Additionally, results show that there is no significant difference between the effectiveness of verification and the effectiveness of Notices in slowing the spread of fake news, and neither feature is perceived as strongly credible or effective.

<p>  </p> <p>Large-scale deep learning models have reached previously unattainable performance for various tasks. However, the ever-growing resource consumption of neural networks generates large carbon footprint, brings difficulty for academics to engage in research and stops emerging economies from enjoying growing Artificial Intelligence (AI) benefits. To further scale AI to bring more benefits, two major challenges need to be solved. Firstly, even though large-scale deep learning models achieved remarkable success, their performance is still not satisfactory when fine-tuning with only a handful of examples, thereby hindering widespread adoption in real-world applications where a large scale of labeled data is difficult to obtain. Secondly, current machine learning models are still mainly designed for tasks in closed environments where testing datasets are highly similar to training datasets. When the deployed datasets have distribution shift relative to collected training data, we generally observe degraded performance of developed models. How to build adaptable models becomes another critical challenge. To address those challenges, in this dissertation, we focus on two topics: few-shot learning and domain adaptation, where few-shot learning aims to learn tasks with limited labeled data and domain adaption address the discrepancy between training data and testing data. In Part 1, we show our few-shot learning studies. The proposed few-shot solutions are built upon large-scale language models with evolutionary explorations from improving supervision signals, incorporating unlabeled data and improving few-shot learning abilities with lightweight fine-tuning design to reduce deployment costs. In Part 2, domain adaptation studies are introduced. We develop a progressive series of domain adaption approaches to transfer knowledge across domains efficiently to handle distribution shifts, including capturing common patterns across domains, adaptation with weak supervision and adaption to thousands of domains with limited labeled data and unlabeled data. </p>

The extensive digital development has fundamentally changed the way we receive information. As a result, new security policy challenges have arisen, and it is therefore important to increase resilience to the dissemination of deceptive information that risks harming society and democracy. The purpose of this study is to investigate which factors covariate with an uncritical approach to misleading information in social media among the Swedish population, with the hope that the results can form the basis for targeted attempts to strengthen resilience. The study examines the specific factors age, trust in traditional news media as well as politicians, and opinion on the factual issue that the misleading information expresses. Using survey data collected by Statistics Sweden and the Swedish Defense Research Agency, the covariation of these factors with the respondents' degree of critical approach to information on social media is analyzed through multiple regression analysis (OLS). The study finds that middle-aged individuals are somewhat more critical of information than other age groups are, especially the older age groups. The differences between the age groups are, however, very small. Individuals who have a high level of trust in traditional news media are, on average, slightly less critical of false information on social media than those who have a lower level of trust. The same applies to individuals who have a high level of trust in politicians. The most prominent finding of the study is that respondents are significantly less critical of false posts when they agree with the opinion expressed in the post. Thus, confirmation bias seems to be of great importance for how we value the credibility of information. This indicates that efforts to convert misconceptions in people who have been misled by false information are ineffective, as individuals have difficulty believing in information that challenges their existing position. Preventive efforts to promote critical thinking in society could therefore be more effective to avoid people getting misled in the future. / Den omfattande digitala utvecklingen har förändrat vårt sätt att ta till oss information i grunden. Till följd av detta har nya säkerhetspolitiska utmaningar uppkommit, och att öka motståndskraften mot spridning av vilseledande information som riskerar att skada samhället och demokratin är därför angeläget. Syftet med denna studie är att undersöka vilka faktorer som samvarierar med ett okritiskt förhållningssätt till vilseledande information i sociala medier hos den svenska befolkningen, med förhoppningen om att resultaten ska kunna ligga till grund för riktade insatser för att stärka motståndskraften mot påverkansförsök. I studien undersöks de specifika faktorerna ålder, förtroende för traditionell nyhetsmedia samt för politiker och åsikt i sakfrågan som den vilseledande informationen uttrycker. Med hjälp av enkätdata insamlad av SCB och FOI analyseras dessa faktorers samvariation med respondenternas grad av kritiskt förhållningssätt till information i sociala medier genom multipel regressionsanalys (OLS). Studien finner att individer i medelåldern förhåller sig något mer kritiskt till information än resterande åldersgrupper, särskilt jämfört med de äldsta respondenterna. Skillnaderna mellan åldersgrupperna är dock väldigt små. Individer som har ett högt förtroende för traditionell nyhetsmedia är i genomsnitt något mindre kritiska till falsk information i sociala medier än de som har ett lägre förtroende. Detsamma gäller för individer som har ett högt förtroende för politiker. Den mest framträdande upptäckten som studien gör är att respondenterna är betydligt mindre kritiska till falska inlägg när de håller med om den åsikt som inlägget uttrycker. Således tycks konfirmeringsbias ha en stor betydelse för hur vi värderar trovärdigheten hos information. Detta indikerar att insatser för att konvertera felaktiga uppfattningar hos personer som har blivit vilseledda av falsk information är ineffektiva, eftersom individer har svårt att tro på information som utmanar deras befintliga ståndpunkt. Insatser som verkar förebyggande för att främja det kritiska tänkandet i samhället skulle därför kunna vara mer verksamma för att förhindra att människor vilseleds i framtiden.

La presente investigación tiene como objetivo analizar la respuesta e interacción de los jóvenes estudiantes de la Facultad de Comunicaciones de la UPC dentro de grupos cerrados de la red social Facebook de la universidad con fake news o noticias falsas acerca del COVID-19. Para lograr el objetivo de la investigación se observa la publicación de noticias falsas y la interacción de los miembros del grupo cerrado de la UPC con respecto a ellas, durante la vigencia de la cuarentena por emergencia sanitaria en el Perú. A partir de los datos recogidos en la observación analizamos la respuesta e interacción de los miembros, utilizando los conceptos de noticia, noticia falsa y pos verdad. / This research aims to analyze the response and interaction of young students from the UPC's Faculty of Communications within closed groups of the university's Facebook social network with fake news about COVID-19. To achieve the objective of the investigation, the publication of fake news and the interaction of the members of the closed group of the UPC with respect to them are observed, during the validity of the quarantine for a health emergency in Peru. From the data collected in the observation, we analyzed the response and interaction of the members, using the concepts of news, false news and post truth. / Trabajo de investigación

The phenomenon of "fake news", or misleading online content, is increasingly worrisome due to its large-scale socio-economic impact. Researchers and practitioners attempted to understand what drives the virality and believability of fake news and how to reduce its influence. This research aims to shed light on these questions. Building upon a theoretical account positing that people share fake news because they simply fail to engage in deliberate thinking, we designed an accuracy prompt intervention to encourage people to think effortfully. In a pre-registered study conducted via Prolific (N = 520), we find limited evidence supporting accuracy prompts stylized as warning labels, but only for increasing sharing discernment in true, not fake news. The veracity of news articles does not impact sharing intentions, despite having a sizeable effect on accuracy judgments. This and other findings support the dual processing theory of cognition in the context of fake news. Predispositions towards more intuitive thinking increased belief in fake news and higher distrust in true news. Conversely, a better ability to engage in effortful thinking increases truth discernment. In addition, confirmation bias decreases truth discernment and increases sharing intentions. Politically concordant true headlines are...

No description available.

No description available.

During the past several decades, journalists have been struggling with the ever growing amount of data on the internet. Investigating the validity of the sources or finding similar articles for a story can consume a lot of time and effort. These issues are even amplified by the declining size of the staff of news agencies. The solution is to empower the remaining professional journalists with digital tools created by computer scientists. This thesis project is inspired by an idea to provide software support for journalistic work with interactive visual interfaces and artificial intelligence. More specifically, within the scope of this thesis project, we created a backend module that supports several text mining methods such as keyword extraction, named entity recognition, sentiment analysis, fake news classification and also data collection from various data sources to help professionals in the field of journalism. To implement our system, first we gathered the requirements from several researchers and practitioners in journalism, media studies, and computer science, then acquired knowledge by reviewing literature on current approaches. Results are evaluated both with quantitative methods such as individual component benchmarks and also with qualitative methods by analyzing the outcomes of the semi-structured interviews with collaborating and external domain experts. Our results show that there is similarity between the domain experts' perceived value and the performance of the components on the individual evaluations. This shows us that there is potential in this research area and future work would be welcomed by the journalistic community.