Aims: The aim was to examine how loneliness was associated with bullying victimization at school and online. Methods: We used data from the Danish arm of the international Health Behaviour in School-aged Children (HBSC) study from 2022. The study population was a nationally representative sample of 11-15-year-olds who completed the internationally standardized HBSC questionnaire at school, n=5382. Multilevel logistic regression was applied to study the associations between bullying victimization and loneliness. Results: The prevalence of reporting loneliness often or very often was 9.0%; 6.3% of the sample experienced habitual bullying victimization at school and 4.8% incurred cyberbullying. There was a strong and graded association between loneliness and bullying victimization at school and cyberbullying. The associations were significant for boys and girls and the association between exposure to bullying at school and loneliness was steeper for boys than girls. The gradients were steeper for physical than cyberbullying. Students exposed to habitual bullying in both contexts had an adjusted OR (95% CI) of 11.21 (6.99-17.98) for loneliness. Conclusion: Exposure to bullying at school and cyberbullying are strongly associated with loneliness. It is important to reduce bullying at school and at the internet, and to promote effective interventions to reduce continuing loneliness.

Many educational institutions are using technology-enhanced learning. However, studies have shown that increased access to technological tools/gadgets is a precursor to incidents of cyberbullying. Using Bronfenbrenner’s Ecological Theory, we explored educators’ knowledge and intervention strategies regarding cyberbullying in schools. The design used in this study was qualitative and descriptive in nature. Purposive sampling technique was used to recruit a sample of eight educators working at secondary school in Johannesburg, South Africa. Data were collected using one-on-one interviews. Thematic analysis method was used to analyse the data. The findings revealed educators had sufficient knowledge of cyberbullying and how it affects learners in schools. However, interventions the teachers had put in place to prevent and manage incidences of cyberbullying were found to be ineffective due to parental disengagement and learners not reporting cases of cyberbullying. Recommendations for enhancing management of cyberbullying cases are provided.

Cyberaggression is an important problem today; it can affect adolescents in different ways. Therefore, reliable and valid measures are necessary to better study the phenomenon. The aim of the present study was to generate validity and reliability evidence for a Spanish-language cyberaggression scale in a population of adolescents from northwestern Mexico. The results of this study contribute to the research and focus on cyberaggression in adolescents in Mexico. The measure in this article differs from other measures by detecting the different roles, including the bystander, contributing to the state of knowledge about cyberaggression, ratifying the presence of this actor that has not been sufficiently explored in the cyber context.

Online social networks (OSNs) play an integral role in facilitating social interaction; however, these social networks increase antisocial behavior, such as cyberbullying, hate speech, and trolling. Aggression or hate speech that takes place through short message service (SMS) or the Internet (e.g., in social media platforms) is known as cyberbullying. Therefore, automatic detection utilizing natural language processing (NLP) is a necessary first step that helps prevent cyberbullying. This research proposes an automatic cyberbullying method to detect aggressive behavior using a consolidated deep learning model. This technique utilizes multichannel deep learning based on three models, namely, the bidirectional gated recurrent unit (BiGRU), transformer block, and convolutional neural network (CNN), to classify Twitter comments into two categories: aggressive and not aggressive. Three well-known hate speech datasets were combined to evaluate the performance of the proposed method. The proposed method achieved promising results. The accuracy of the proposed method was approximately 88%.

The present work analyzes university students' perceptions on cyberbullying. More specifically, the main objective was to understand the level of concern displayed by young students in teacher training programs regarding cyberbullying; our analysis includes their degree of self-confidence and their commitment when faced with this problem, their self-reported training on the subject, and the measures they consider adequate to approach it in the educational context. Using a survey research design, 408 students from Early Childhood Education, Primary Education Degrees, and the Master’s Degree in Secondary Education, and Vocational Training from national public universities participated. The results shed light upon the coexistence of three differentiated profiles of teachers in training when faced with cyberbullying. However, although there is clear evidence of the high level of concern regarding this problem in every case, they placed more importance on raising awareness of the issue with families, prevention, management and training as basic pillars for re-directing the worrisome reality experienced in educational centers. The need to include these contents in the initial training of future education professionals, for overcoming cyberbullying, and the importance of collaborative work between parents and educators are the conclusions of this research study that are in agreement with other studies.

Literature indicates scarcity of cross-national research on the association between physical fighting and bullying victimisation among adolescents, and the moderating role of perceived social support and education spending on this association. Using multilevel binary logistic regression, this study examined the association of physical fighting with traditional bullying and cyberbullying victimisation, with public education spending and perceived social support from families and teachers as moderators. Country-level data were combined with 2017/18 Health Behaviour in School-aged Children (HBSC) survey data from 162,792 adolescents (11-, 13- and 15-year-olds) in 27 European countries. Results showed that physical fighting was positively and significantly associated with cyberbullying and traditional bullying victimisation. These results signify that physical fighting is a robust predictor of bullying victimisation. The findings demonstrate that perceived social support and education spending buffered the association of physical fighting with traditional bullying and cyberbullying victimisation. The findings imply that social support not only prevents adolescents from engaging in physical fights, but also reduces the risk of bullying victimisation. Accordingly, to deter adolescents from engaging in physical violence and reduce the risk of bullying victimisation, there is need to fully comprehend the influence of physical fighting on bullying victimisation, and the moderating role of social support.

The dominance of social media has added to the channels of bullying to perpetrators. Unfortunately, cyberbullying (CB) is the most prevalent phenomenon in today’s cyber world and is a severe threat to the mental and physical health of citizens. This opens the need to develop a robust system to prevent bullying content from online forums, blogs, and social media platforms to manage the impact in our society. Several machine learning (ML) algorithms have been proposed for this purpose. However, their performances are not consistent due to high class imbalance issue and generalisation. In recent years, large language models (LLM) like BERT and RoBERTa have achieved state of the art (SOTA) results in several natural language processing (NLP) tasks. Unfortunately, the LLMs have not been applied extensively. In our paper, we explored the use of these models for cyberbullying (CB) detection. We have prepared a new dataset (D2) from existing studies (Formspring and Twitter). Our experimental results for dataset D1 and D2 showed RoBERTa outperformed other models.

(1) Background: Violence against each other via social media has increased and caused cyberbullying that can happen anytime through electronic communication tools that everyone can access easily. So, this study aims to investigate causal factors contributing to youth cyberbullying in Thailand's three Southern border provinces. (2) Methods: The subject group consisted of 340 youths and the data were analyzed with structural equation modeling (SEM). (3) Results: The study found that the model of the causal factors resulting in youth cyberbullying in the three Southern border provinces was consistent with the empirical data (the relative chi-square (χ2/df) was 1.77. The goodness-of-fit index (GFI) was 0.95. The root-mean-square-error of approximate (RMSEA) was 0.049). Cyberbullying was positively influenced directly and indirectly by negative upbringing, the influence of personal violence, and the influence of media violence at a statistically significant level of 0.001, with total effect sizes of 1.13, 0.74, and 0.64, respectively. Additionally, cyberbullying was positively influenced directly by negative mental traits with a statistically significant level of 0.05 and a total effect size of 0.17. (4) Conclusions: The study results could be beneficial in concretely forming policies and strategies to prevent and mitigate the problem of youth cyberbullying.

Despite significant advancements in Artificial Intelligence (AI) and Large Language Models (LLMs), detecting and mitigating bias remains a critical challenge, particularly within social media platforms like X (formerly Twitter) and in addressing cyberbullying present on them. This research investigates the effectiveness of leading LLMs in generating synthetic biased and cyberbullying data and evaluates the proficiency of Transformer AI models in detecting bias and cyberbullying within both authentic and synthetic contexts. The study involves semantic analysis and feature engineering on a dataset of over 48,000 sentences related to cyberbullying collected from Twitter (before it became X). Leveraging state-of-the-art LLMs such as ChatGPT-4o, Pi AI, Claude 3 Opus, and Gemini-1.5, synthetic biased, cyberbullying, and neutral data were generated to deepen the understanding of bias in human-generated data. AI models including DeBERTa, Longformer, BigBird, HateBERT, MobileBERT, DistilBERT, BERT, RoBERTa, ELECTRA, and XLNet were initially trained to classify Twitter cyberbullying data and subsequently fine-tuned, optimized, and quantized for multilabel classification (detecting both biases and cyberbullying). The study's outcomes include a prototype of a hybrid application that combines a Bias Data Detector and a Bias Data Generator, validated through extensive testing.