Search Results (322)

In an era where online interactions increasingly shape social dynamics, the pervasive issue of cyberbullying poses a significant threat to the well-being of individuals, particularly among vulnerable groups. Despite extensive research on text-based cyberbullying detection, the rise of visual content on social media platforms necessitates new approaches to address cyberbullying using images. This domain has been largely overlooked. In this paper, we present a novel dataset specifically designed for the detection of visual cyberbullying, encompassing four distinct classes: abuse, curse, discourage, and threat. The initial prepared dataset (cyberbullying visual indicators dataset (CVID)) comprised 664 samples for training and validation, expanded through data augmentation techniques to ensure balanced and accurate results across all classes. We analyzed this dataset using several advanced deep learning models, including VGG16, VGG19, MobileNetV2, and Vision Transformer. The proposed model, based on DenseNet201, achieved the highest test accuracy of 99%, demonstrating its efficacy in identifying the visual cues associated with cyberbullying. To prove the proposed model’s generalizability, the 5-fold stratified K-fold was also considered, and the model achieved an average test accuracy of 99%. This work introduces a dataset and highlights the potential of leveraging deep learning models to address the multifaceted challenges of detecting cyberbullying in visual content. Full article

Social media has become a platform for the rapid spread of persuasion techniques that can negatively affect individuals and society. Propaganda detection, a crucial task in natural language processing, aims to identify manipulative content in texts, particularly in news media, by assessing propagandistic intent. Although extensively studied in English, Arabic propaganda detection remains challenging because of the language’s morphological complexity and limited resources. Furthermore, most research has treated propaganda detection as an isolated task, neglecting the influence of sentiments and emotions. The current study addresses this gap by introducing the first multi-task learning (MTL) models for Arabic propaganda detection, integrating sentiment analysis and emotion detection as auxiliary tasks. Three MTL models are introduced: (1) MTL combining all tasks, (2) PSMTL (propaganda and sentiment), and (3) PEMTL (propaganda and emotion) based on transformer architectures. Additionally, seven task-weighting schemes are proposed and evaluated. Experiments demonstrated the superiority of our framework over state-of-the-art methods, achieving a Macro-F1 score of 0.778 and 79% accuracy. The results highlight the importance of integrating sentiment and emotion for enhanced propaganda detection; demonstrate that MTL improves model performance; and provide valuable insights into the interaction among sentiment, emotion, and propaganda. Full article

The proliferation of social media platforms has triggered an unprecedented increase in multimodal fake news, creating pressing challenges for content authenticity verification. Current fake news detection systems predominantly rely on isolated unimodal analysis (text or image), failing to exploit critical cross-modal correlations or leverage latent social context cues. To bridge this gap, we introduce the SCCN (Semantic-enhanced Cross-modal Co-attention Network), a novel framework that synergistically combines multimodal features with refined social graph signals. Our approach innovatively combines text, image, and social relation features through a hierarchical fusion framework. First, we extract modality-specific features and enhance semantics by identifying entities in both text and visual data. Second, an improved co-attention mechanism selectively integrates social relations while removing irrelevant connections to reduce noise and exploring latent informative links. Finally, the model is optimized via cross-entropy loss with entropy minimization. Experimental results for benchmark datasets (PHEME and Weibo) show that SCCN consistently outperforms existing approaches, achieving relative accuracy enhancements of 1.7% and 1.6% over the best-performing baseline methods in each dataset. Full article

To obtain actionable information for humanitarian and other emergency responses, an accurate classification of news or events is critical. Daily news and social media are hard to classify based on conveyed information, especially when multiple categories of information are embedded. This research used large language models (LLMs) and traditional transformer-based models, such as BERT, to classify news and social media events using the example of the Sudan Conflict. A systematic evaluation framework was introduced to test GPT models using Zero-Shot prompting, Retrieval-Augmented Generation (RAG), and RAG with In-Context Learning (ICL) against standard and hyperparameter-tuned bert-based and bert-large models. BERT outperformed GPT in F1-score and accuracy for multi-label classification (MLC) while GPT outperformed BERT in accuracy for Single-Label classification from Multi-Label Ground Truth (SL-MLG). The results illustrate that a larger model size improves classification accuracy for both BERT and GPT, while BERT benefits from hyperparameter tuning and GPT benefits from its enhanced contextual comprehension capabilities. By addressing challenges such as overlapping semantic categories, task-specific adaptation, and a limited dataset, this study provides a deeper understanding of LLMs’ applicability in constrained, real-world scenarios, particularly in highlighting the potential for integrating NLP with other applications such as GIS in future conflict analyses. Full article

The complexity for law enforcement and child protection agencies has been exacerbated by the proliferation of child sexual exploitation channels, facilitated by digital platforms and social media. Generative AI’s ability to analyze large datasets, recognize patterns, and generate new content makes it one of the potential solutions for detecting suspicious behavior and indicators of child sexual exploitation. This paper discusses the potential of generative AI to aid in the fight against pedophilic crimes by reviewing current research, methodologies, and challenges, as well as future directions and ethical concerns. Although the potential benefits are significant, applying AI to such a sensitive area presents numerous challenges, including privacy concerns, algorithmic bias, and potential misuse, which must be addressed carefully. Full article

Technological developments and the integration of social media into journalistic practices have transformed the media landscape, changing how information is gathered, produced, and shared. This evolution poses challenges, including the lack of clear guidelines and practical tools for ensuring the quality of digital news content. To address these issues, IQJournalism, an intelligent quality prediction advisor, was developed. This paper outlines the methodology for the development of IQJournalism, a platform that leverages advanced AI technologies to process Greek news articles and provide real-time editing recommendations on various dimensions, including language quality, subjectivity level, emotionality, entertainment, and social media engagement. First, a qualitative study was conducted through semi-structured, in-depth interviews with 20 experts, academic researchers and media professionals to identify indicators of perceived quality in journalism. These insights were then transformed into measurable features, which served as training data for explainable machine learning-based models for quality categorization and prediction. Finally, the IQJournalism platform was designed following a user-centered iterative process that included prototyping, testing, and redesigning. The innovative approach aims to serve as a valuable tool for improving journalistic quality, contributing to more reliable and engaging online news content. Importantly, the platform is not limited to the journalistic sector, but can also be used to optimize content in various areas, such as marketing, political, and strategic communication, supporting editors seeking to improve the quality and impact of their writing. Full article

It is a popular assumption that people learn certain practices for handling media in the course of their adolescence and adulthood, which make it difficult for them to develop new patterns for the use of media at a later point in their lives. From this theoretical standpoint, it is a challenge for older people to learn how to handle new media and integrate them into their current living situation. Beyond theoretical assumptions, there has formerly been a lack of exploratory investigations pursuing the conditions under which older adults take up digital media with which they were previously not familiar and incorporate them into their daily lives. Between October 2023 and March 2024, 32 semi-standardised individual interviews were conducted with a group of people between 80 and 93 years of age, who had recently acquired a digital medium and integrated it into their everyday lives. The decisive factor here was the presence of certain motives that generate plausible incentives to make permanent use of new media. The interviewees have purposefully acquired new media. It is notable that acquisition processes were strongly initiated by significant changes in life circumstances. In the case of most interviewees, the intention to acquire an internet-enabled medium was based on the wish to use a few selected functions. New options for online use were only explored after a while. The following patterns were identified regarding the motives and gratifications of acquisition: new media as…(1) hobby extension, (2) support network, (3) compensation tool, (4) connection opportunity, (5) escape from everyday life. It can be assumed that older people experience the use of new media as purposeful if they have specific motives for doing so. Biological, psychological and social correlations as well as ways of coping and dealing with age(ing) are relevant here. If daily use potentials are perceived as beneficial, older people show a high level of adaptability in terms of new media. Against this background, a gratification-orientated model appears to be a promising starting point for explaining the prerequisites for media adoption based on motives that generate plausible incentives for learning how to use new media at an older age. Full article

Urbanization is reshaping landscapes and posing unprecedented sustainability challenges, necessitating more integrative approaches to urban observation. This review synthesizes recent advancements in traditional remote sensing and emerging social sensing technologies, emphasizing their convergence within urban science. A systematic thematic analysis of 667 peer-reviewed articles highlights the methodological progress, practical applications, and theoretical innovations arising from this integration. Traditional remote sensing effectively captures urban physical features but lacks insights into human behaviors. Conversely, social sensing, leveraging digital traces from social media and mobile data, introduces essential human-centered dimensions into urban monitoring. The fusion of these complementary paradigms through advanced data analytics and multimodal integration has produced transformative methodologies, enhancing urban resilience frameworks, functional zone delineation, and real-time disaster responses. Despite significant progress, the integration faces persistent challenges, including data heterogeneity, representational bias, ethical concerns, and scalability limitations. Differing from previous reviews that survey the landscape, the current work argues that current integration efforts remain ad hoc and technologically driven, lacking a unifying theory for real-time urban governance. To address this critical gap, I develop and operationalize a new systems-based framework for hybrid urban observatories. This framework is built on a socio-ecological foundation and explicitly integrates technical components with an essential governance layer, advancing both methodological rigor and actionable guidance for the field. Such a framework will enable a more holistic, responsive, and equitable approach to urban governance and sustainability. Full article

The growing use of social media data has opened new avenues for understanding user perceptions and operational inefficiencies in transportation systems. Among the most widely adopted analytical approaches for extracting insights from these data are sentiment analysis and topic modeling, which enable researchers to capture public opinion trends and uncover latent themes in unstructured content. However, despite a rising number of individual studies, systematic reviews focusing specifically on these approaches in transportation research remain limited, particularly in addressing methodological challenges and data heterogeneity. This literature review addresses that gap by critically examining 81 open-access studies published between 2014 and 2024. The main challenges identified include handling linguistic diversity, integrating multimodal and geolocated data, managing short-text formats, and addressing regional and demographic bias. In response, this review proposes a methodological framework for study selection and bibliometric analysis, classifies the most commonly applied machine learning models for sentiment and topic extraction, and synthesizes findings regarding data sources, model performance, and application contexts in transportation. Additionally, it discusses unresolved gaps and ethical concerns related to representativeness and social media governance. This review highlights the transformative potential of combining sentiment analysis and topic modeling to support smarter, more inclusive, and sustainable transportation policies by offering an integrative and critical perspective. Full article

In an age of AI-generated content and deepfakes, fake news and disinformation are increasingly spread using manipulated or fabricated images. To address this challenge, we introduce Post-Quantum VerITAS, a cryptographic framework for verifying the authenticity and history of digital images—even in a future where quantum computers threaten classical encryption. Our system supports common image edits, like cropping or resizing, while proving that the image is derived from a legitimate, signed source. Using quantum-resistant tools, like lattice-based hashing, modified Poseidon functions, and zk-SNARK proofs, we ensure fast, privacy-preserving verification without relying on trusted third parties. Post-Quantum VerITAS offers a scalable, post-quantum-ready solution for image integrity, with direct applications in journalism, social media, and secure digital communication. Full article

This article explores the challenges surrounding the Scottish media graduate landscape after the COVID-19 pandemic. Contributing factors that impact Scotland-based students and educators include a shift in the jobs market, altering pedagogies during and post-pandemic, and social drivers including fewer students choosing media pathways of study due to the cost-of-living crisis. This study draws on insights from 40 students at five Scottish universities, all of whom graduated in the summer of 2023. The research presents a window into the mindset and expectations of this post-pandemic graduating class while drawing on current and relevant literature. In addition, the paper includes reaction from industry and academic experts in Scotland and questions what can be done to address trends surrounding the stability and sustainability of journalism education. The experts include senior broadcasters, an established media educator who has worked across further education and higher education in Scotland while also being a national news editor, and one of Scotland’s most experienced journalism educators who is the chair of the World Journalism Education Council. This work is predominantly qualitative, drawing on a mixed research approach of expert interviewing and surveys while providing recommendations for journalism educators. Full article

Political communication has been extensively studied through both the broader context of societal and political systems, as well as through the lens of mediatization, which emphasizes the intersection of political and media logics. Within this framework, scholars originally identified three distinct stages or eras of political communication. However, recent scholarship has increasingly focused on the transition to a “fourth” era, characterized by the growing impact of digital and social media. This shift, from the “television age” to the “social media age”, has not only introduced new media channels for conveying political messages but has also fundamentally transformed the nature of political communication itself—shifting from top-down, centralized models to more horizontal, decentralized forms of interaction. Current research on the role of social media in political communication reveals a complex landscape. These platforms appear to both enhance and undermine established processes of political deliberation. On the one hand, they provide new avenues for civic engagement and political discourse, while on the other, they contribute to issues such as disinformation, polarization, and the erosion of privacy. This entry aims to offer a comprehensive review of how social media platforms have reshaped the dynamics of political communication and civic participation. It further explores the challenges that accompany these transformations, such as the spread of disinformation, rising political polarization, increasing incivility, and privacy concerns stemming from advanced digital marketing techniques in political contexts. Full article

The evolution of social bots, i.e., accounts on social media platforms controlled by malicious software, is making them increasingly more challenging to discover. A practical solution is to explore the adversarial nature of novel bots and find the vulnerability of bot detectors in simulations in advance. However, current studies fail to realistically simulate the environment and bots’ actions, thus not effectively representing the competition between novel bots and bot detectors. Hence, we propose a new method for modeling the impact of bot actions and develop a new bot strategy to simulate various evolved bots within a large social network. Specifically, a bot influence model and a user engagement model are introduced to simulate the growth of followers, retweets, and mentions. Additionally, a profile editor and a target preselection mechanism are proposed to more accurately simulate the behavior of evolved bots. The effectiveness of the bots and two representative bot detectors are verified using adversarial simulations and the real-world dataset. In simulated adversarial scenarios against both RF-based and GNN-based detection models, the proposed approach achieves survival rates of 99.7% and 85.9%, respectively. The simulation results indicate that, despite utilizing the bots’ profile data, user-generated content, and graph information, the detectors failed to identify all variations of the bots and mitigate their impact. More importantly, for the first time, it is found that certain types of bots outperform those usually deemed more advanced in ablation experiments, demonstrating that such “penetration testing” can indeed reveal vulnerabilities in the detectors. Full article

This study examines the evolving relationship between online media coverage and protest movements by analyzing year-long demonstrations in Israel against Prime Minister Benjamin Netanyahu. Through comprehensive qualitative thematic analysis and content analyses of 219 online newspaper articles from five major Israeli newspapers; 324 social media posts across Facebook, Instagram, and Twitter; and 9 semi-structured interviews with protest leaders, this research identifies a gradual paradigm shift in protest representation in online media. The findings reveal a transition from the traditional “protest paradigm”—which frames protests as violent and remote through warlike discourse and visual distancing—toward an emerging “our protest paradigm”, characterized by rhetorical and visual proximity to protesters. This new paradigm manifests through personal testimonies in mainstream media and portrait photography on social media platforms, both creating a sense of closeness and accountability. The study further reveals a significant disconnect between protest leaders’ perceptions and legacy media, as leaders increasingly view traditional media as irrelevant despite their advisers’ recommendations to engage with it. Using polysystem theory as a theoretical framework, this research demonstrates how two media systems—legacy media and social media—operate with epistemological rigidity, challenging the previously established notion of “competitive symbiosis” between protesters and journalists. This investigation offers a novel analytical perspective through the lens of distance, illuminating how changing dynamics in online information transfer are reshaping protest coverage and production. The resulting paradigm model explains the coexistence of two simultaneous protest paradigms and provides valuable insights into the contemporary relationship between social movements, legacy media, and digital platforms in an evolving media ecosystem. Full article

Social media addiction has become a global public health challenge, and understanding its mechanism’s complexity requires the integration of the transitional characteristics of addiction development stages and breaking through the traditional single-reinforcement-path explanatory framework. This study is based on the dual pathway of positive and negative emotional reinforcement, integrating multidisciplinary evidence from neuroscience, psychology, and computational behavioral science to propose an independent and dynamic interaction mechanism of positive reinforcement (driven by social rewards) and negative reinforcement (driven by emotional avoidance) in social media addiction. Through a review, it was found that early addiction is mediated by the midbrain limbic dopamine system due to immediate pleasurable experiences (such as liking), while late addiction is maintained by negative emotional cycles due to the dysfunction of the prefrontal limbic circuit. The transition from early addiction to late addiction is characterized by independence and interactivity. Based on this, a phased intervention strategy is proposed, which uses reward competition strategies (such as cognitive behavioral therapy and alternative rewards) to weaken dopamine sensitization in the positive reinforcement stage, enhances self-control by blocking emotional escape (such as through mindfulness training and algorithm innovation) in the negative reinforcement stage, and uses cross-pathway joint intervention in the interaction stage. This study provides a theoretical integration framework for interdisciplinary research on social media addiction from a dynamic perspective for the first time. It is recommended that emotional reinforcement variables are included in addiction diagnosis, opening up new paths for precise intervention in different stages of social media addiction development. Full article