The power of rumour spreading in the age of online social media is intimidating. It can incite to insurrection, denigrate people, and damage financial markets, proving catastrophic for society. Despite widespread scholarly research and practice of developing a constellation of counter-rumour strategies, the massive waves of rumours are still sweeping over individuals, organisations, and societal institutions. To systematically tackle this issue, we present a comprehensive review and an epidemic framework to resolve three challenging aspects of rumour dissemination in online social media. First, we identify and explain the various forms of false and unverified information, relevance, and impact. Second, we address how social media can exacerbate the phenomenon of rumour spreading. Using the framework, the classification of rumour disseminating mechanisms on social media, allows us to develop counter-rumour strategies. Finally, we inspect past strategies employed in addressing rumour dissemination and use the framework to explore parallels between epidemic management and addressing rumour. We identify the highly neglected aspects of the current cumulative rumour response and factors that may be effectively targeted in the future. Our approach might support understanding social media’s role in propagating rumours and devising active measures in quelling this epidemic.@en

The phenomenon of rumour spreading refers to a collective process where people participate in the transmission of unverified and relevant information to make sense of the ambiguous, dangerous, or threatening situation. The dissemination of rumours on a large scale no matter with what purpose could precipitate catastrophic repercussions. This research aims at addressing this challenge systematically. More in detail, the primary research objective of this dissertation is To systematically study the rumour confrontation within online social media. To accomplish this objective, six steps are taken. At first, the conceptualisation of the main construct in this research is investigated. There are myriad of concepts in English language implying false or unverified information. However, despite years of academic research, there is no consensus regarding their conceptualisation, and they are often used interchangeably or conflated into one idea. This problem could become an obstacle to countering the surge of false information by creating confusion, distracting the community’s attention, and draining their efforts. In the first step, this dissertation addresses this challenge by providing a process-based reading of false and unverified information. This view argues that although the genesis of such information might be deliberate or inadvertent and with different purposes, they primarily disseminate on the basis of similar motives and follow the same process. After settling the conceptualisation problem, the next step investigates the role of communication mediums and especially online social media in the spread of rumours. Although the phenomenon of rumour dissemination has drawn much attention over the past few years, it is an ancient phenomenon. The rumours used to circulate through primitive forms of communications such as word of mouth or letters; however, the technological development, particularly social media, escalated the scale, speed, and scope of this phenomenon. This step aims to pinpoint the features privy to social media that facilitate the emergence and the spread of rumours. Especially, an exclusive automation mechanism of recommendation systems in social media is closely examined through a set of experiments based on YouTube data. The third step in this study investigates the constellation of past counter-rumour strategies. Although rumour spreading and its potentially destructive effects have been taken into account since ancient times, it was only less than a century ago that the first systematic efforts against the mass spread of rumours began. Since then, a series of strategies have been practised by various entities; nevertheless, the massive waves of rumours are still sweeping over individuals, organisations, and societal institutions. In order to develop an effective and comprehensive plan to quell rumours, it is crucial to be aware of the past counter strategies and their potential capabilities, shortcomings and flaws. In this step, we collect the counter strategies over the past century and set them in the epidemic control framework. This framework helps to analyse the purpose of the strategies which could be (i) exposure minimisation, (ii) immunisation or vaccination, and (iii) reducing the transmission rate. The result of the analysis allows us to understand, what aspects of confrontation with rumour have been targeted extensively and what aspects are highly neglected. Following the discussion on the epidemic framework, one of the most effective approaches to rumour confrontation is the immunisation which is primarily driven by academia. The fourth step investigates the readiness of academia in this subject domain. When we do not know the readiness level in a particular subject, we either overestimate or underestimate our ability in that subject. Both of these misjudgements are incorrect and lead to decisions irrelevant to the existing circumstance. To tackle this challenge, the technology emergence framework is deployed to measure academia's readiness level in the topic of rumour circulation. In this framework, we study four dimensions of emergence (novelty, growth, coherence and impact) over more than 21,000 scientific articles, to see the level of readiness in each dimension. The results show an organic growth which is not sufficiently promising due to the surge of rumours in social media. This challenge could be tackled by creating exclusive venues that lead to the formation of a stable community and realisation of an active field for rumour studies. The other aspect of the epidemic framework involves exposure minimisation and transmission rate reduction, which are addressed in the fifth step by an artificial intelligence based solution. The drastic increase in the volume, velocity, and the variety of rumours entails automated solutions for the inspection of circulating contents in social media. In this vein, binary classification is a dominant computational approach; however, it suffers from non-rumour pitfall, which makes the classifier unreliable and inconsistent. To address this issue a novel classification approach is utilised which only uses one rather than multiple classes for the training phase. The experimentation of this approach on two major datasets shows a promising classifier that can recognise rumours with a high level of F1-score. The last step of this manuscript approaches the topic of rumour confrontation from a pro-active perspective. The epidemic framework helps to develop solutions to control rumour dissemination; however, they mostly adopt a passive approach which is reactive and after-the-fact. This step introduces an ontology model that can capture the underlying mechanisms of social manipulation operations. This model takes a proactive stance against social manipulation and provides us with an opportunity of developing preemptive measures. The model is evaluated by the experts and through exemplification on three notoriously famous social manipulation campaigns.@en

The spread of misinformation is one of the severe challenges that societies have been dealing with for many years. However, the rapid growth of social media has accelerated the creation and circulation of such information and turned it into a potential threat to the main societal institutions such as peace and democracy. Although many of iconic figures, policymakers, business leaders and researchers have warned us of serious repercussions of misinformation, a clear course of action is not yet visible. To tackle such an issue, the preliminary step would be the evaluation of the as-is situation, which allows us to identify the deficiencies of existing solutions. This issue has been addressed in this study by a comprehensive analysis over decades of societal efforts against misinformation. In this analysis, quelling strategies from organisational and government perspectives are explained. Then they are investigated from efficacy level and governance mode. Our analyses show that, despite a seemingly suitable setting for confronting misinformation, there is a major shortcoming in governance mode of current quelling strategies.

@en

The online spread of rumours in disasters can create panic and anxiety and disrupt crisis operations. Hence, it is crucial to take measure against such a distressing phenomenon since it can turn into a crisis by itself. In this work, the automatic rumour detection in natural disasters is addressed from an imbalanced learning perspective due to the rumour dearth versus non-rumour abundance in social networks. We first provide two datasets by collecting and annotating tweets regarding the Hurricane Florence and Kerala flood. We then capture the properties of rumours and non-rumours in those disasters using 83 theory-based and early-available features, 47 of which are proposed for the first time. The proposed features show a high discrimination power that help us distinguish rumours from non-rumours more reliably. Next, We build the rumour identification models using imbalanced learning to address the scarcity of rumours compared to non-rumour. Additionally, to replicate the rumour detection in the real-world situation, we practice cross-incident learning by training the classifier with the samples of one incident and test it with the other one. In the end we measure the impact of imbalanced learning using Bayesian Wilcoxon Signed-rank test and observe a significant improvement in the classifiers performance.

@en

YouTube has been implicated in the transformation of users into extremists and conspiracy theorists. The alleged mechanism for this radicalizing process is YouTube’s recommender system, which is optimized to amplify and promote clips that users are likely to watch through to the end. YouTube optimizes for watch-through for economic reasons: people who watch a video through to the end are likely to then watch the next recommended video as well, which means that more advertisements can be served to them. This is a seemingly innocuous design choice, but it has a troubling side-effect. Critics of YouTube have alleged that the recommender system tends to recommend extremist content and conspiracy theories, as such videos are especially likely to capture and keep users’ attention. To date, the problem of radicalization via the YouTube recommender system has been a matter of speculation. The current study represents the first systematic, pre-registered attempt to establish whether and to what extent the recommender system tends to promote such content. We begin by contextualizing our study in the framework of technological seduction. Next, we explain our methodology. After that, we present our results, which are consistent with the radicalization hypothesis. Finally, we discuss our findings, as well as directions for future research and recommendations for users, industry, and policy-makers.@en

Rumour is a collective emergent phenomenon with a potential for provoking a crisis. Modelling approaches have been deployed since five decades ago; however, the focus was mostly on epidemic behaviour of the rumours which does not take into account the differences between agents. We use social practice theory to model agent decision-making in organizational rumourmongering. Such an approach provides us with an opportunity to model rumourmongering agents with a layer of cognitive realism and study the impacts of various intervention strategies for prevention and control of rumours in organizations.

@en

The spread of false and unverified information has the potential to inflict damage by harming the reputation of individuals or organisations, shaking financial markets, and influencing crowd decisions in important events.This phenomenon needs to be properly curbed, otherwise it can contaminate other aspects of our social life. In this regard, academia as a key institution against false and unverified information is expected to play a pivotal role. Despite a great deal of research in this arena, the amount of progress by academia is not clear yet. This can lead to misjudgements about the performance of the topic of interest that can ultimately result in wrong science policies regarding academic efforts for quelling false and unverified information. In this research, we address this issue by assessing the readiness of academia in the topic of false and unverified information. To this end, we adopt the emergence framework and measure its dimensions (novelty, growth, coherence, and impact) over more than 21,000 articles published by academia about false and unverified information. Our results show the current body of research has had organic growth so far, which is not promising enough for confronting the problem of false and unverified information. To tackle this problem, we suggest an external push strategy that, compared to the early stages of the topic of interest, reinforces the emergence dimensions and leads to a higher level in every dimension.@en

This study addresses the problem of rumour scarcity versus non-rumour abundance in automatic rumour detection. To tackle this issue, we portray rumour as an anomaly by showing how disproportionate is the number of rumours versus non-rumours. This imbalance is scrutinized by comparing the rate of news production versus rate of fact-check production. Then, we exploit one-class classification approach to distinguish rumour from non-rumour. One-class classification separates rumour from non-rumour via training the classifier with only non-rumour. To train the one-class classifier, we extract 33 short-term features, regarding the purpose of this research in early detection of rumours. We evaluate the performance of our model by accuracy and F-score. In terms of F-score, our model outperforms the state-of-the-art and reaches to very close proximity of highest accuracy on the same dataset.@en

Rumor spreading in online social networks can inflict serious damages on individual, organizational, and societal levels. This problem has been addressed via computational approach in recent years. The dominant computational technique for the identification of rumors is the binary classification that uses rumor and non-rumor for the training. In this method, the way of annotating training data points determines how each class is defined for the classifier. Unlike rumor samples that often are annotated similarly, non-rumors get their labels arbitrarily based on annotators' volition. Such an approach leads to unreliable classifiers that cannot distinguish rumor from non-rumor consistently. In this paper, we tackle this problem via a novel classification approach called one-class classification (OCC). In this approach, the classifier is trained with only rumors, which means that we do not need the non-rumor data points at all. For this study, we use two primary Twitter data sets in this field and extract 86 features from each tweet. We then apply seven one-class classifiers from three different paradigms and compare their performance. Our results show that this approach can recognize rumors with a high level of F1-score. This approach may influence the predominant mentality of scholars about computational rumor detection and puts forward a new research path toward dealing with this problem.@en

For the period surrounding the 2018 Dutch municipal elections, a team of researchers from the Delft University of Technology investigated the effect of the digital environment on parliamentary democracy. An interdisciplinary group of researchers combined expertise on digital ethics, political theory, big data analytics, the economics of privacy and security, epistemology, media studies and computer science. This report presents the main findings, which are grouped around two main themes: political micro-targeting and ICT media. Societal themes that came to prominence over the research period, such as the debate over ‘fake news’ and the leaks of personal information that were used for political purposes by Facebook, as well as the implementation of new EU privacy regulation helped to put the research in a larger political context. The main findings provide a qualified picture. The influence of the digital revolution on democratic politics is already revolutionary, and the weaknesses of online platforms provide ample opportunities for derailing liberal democracy. Digital platforms are too closed-off, not mindful enough of individual digital rights, and biased in their (re)presentation of political pluralism. But the Netherlands has proven to be one of the few democracies that is relatively resilient, with an open multi-party system receptive to the political fragmentation that ICT developments encourage, and relatively high trust between citizens, in shared media organizations, and between political parties. In order not to be complacent in the face of fundamental challenges, the report provides several urgent recommendations. Next to several ‘reactive’ recommendations, which seek to remedy the weaknesses and dangers the digital environment poses to democracy, it also outlines an example of how the digital environment might be proactively redesigned in order to positively enhance the quality of the Dutch parliamentary system. @en

On August 25, 2017, Hurricane Harvey made landfall near Rockport, Texas as a Category 4 hurricane with maximum sustained winds of approximately 200 km/hour. Harvey caused severe damages in coastal Texas due to extreme winds and storm surge, but will go down in history for record-setting rainfall totals and flood-related damages. Across large portions of southeast Texas, rainfall totals during the six-day period between August 25 and 31, 2017 were amongst the highest ever recorded, causing flooding at an unprecedented scale. More than 100,000 residential properties are estimated to have been affected in southeast Texas. It is likely that Harvey will rank among the costliest storms in U.S. history. In the wake of Hurricane Harvey, Delft University of Technology has initiated a Harvey Research Team to undertake a coordinated multidisciplinary investigation of the events with a focus on the greater Houston area. This ‘fact-finding’ research is based on information available from public sources during and in the first weeks after the event. Results are therefore preliminary, but aim to provide insight into lessons that can be learned for both Texas and the Netherlands. As part of the investigations, a hackathon with more than 80 participants was organized to collect and analyze available public information. Houston was especially hard hit by flooding. During the event, all 22 watersheds in the greater Houston area experienced flooding. Many of Houston’s creeks and bayous exceeded their channel capacities, reaching water levels never before recorded. Across large portions of Harris County, rainfall totals exceeded the 1000-year return period. In addition, the water from the two reservoirs protecting downtown Houston (Addicks and Barker) were opened on August 28 to prevent catastrophic damages to the dams and further flooding in upstream communities. The releases exacerbated flooding in the areas downstream of the dams and an estimated 4,000 homes in neighborhoods downstream of the dams were impacted by flooding. The consequences of the event in the greater Houston area have been characterized in terms of economic damages, loss of life and impacts on critical infrastructure, airports and industry. In total, more than 100,000 homes were affected more than 70 fatalities were reported in the greater Houston area. The event highlighted the vulnerability of industrial facilities, as several cascading impacts (releases of toxic materials and explosions) were reported. Emergency response has been assessed. No large-scale mandatory evacuation was ordered before or during Harvey. However, it appeared that several local evacuations were ordered for areas with specific risks and circumstances. During the event, many people were trapped by rising waters necessitating a major rescue operation. In total, more than 10,000 rescues were made by professional and volunteer rescuers. Social media played an important role during the event and recovery, as an additional source of information, to inform emergency managers and as a means to organize community response e.g. for clean-up. Also, messages were conveyed through social media, e.g. a report of a levee breach that appeared to be incorrect afterwards. Major flooding is a problem that has multiple causes from both physical and social origin. Based on the investigations, recommendations for future research and lessons for flood management have been formulated. A better understanding of the issues studied in this report is expected to contribute to a knowledge basis for further in-depth investigations and future directions for flood risk reduction. Data collection and Report production funded by DIMI and DSys Special Case 'Houston Galveston Bay Region, Texas, USA' Project 'Harvey hackathon' and follow-up research @en