Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{190525,
        author = {Meenakshi M and Jyothika Sreevalsan and Swathi K Santhosh and Viswajith C P and Rebitha K R},
        title = {Cyberbullying Detection Using Artificial Intelligence: A Survey of Methods and Challenges},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {8},
        pages = {4471-4475},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=190525},
        abstract = {Cyberbullying is a significant issue on social media platforms that is impacting the mental health and well-being of many people and students, as well as their social relationships. With the increasing amount of online communication occurring each day between people and groups, the current methods of moderating and reporting bullying must evolve to become automated so that cyberbullying can be identified and stopped more efficiently.
This survey presents a summary and organization of the available methods used by AI and ML for the detection and prevention of cyberbullying. Additionally, it analyzes the previously developed studies by reviewing the types of techniques, data, feature extraction methods, optimization strategies and evaluation metrics used by researchers.
The papers in this survey were collected from peer-reviewed publications in major digital libraries. The focus was on recent studies about cyberbullying detection. We analyzed and categorized the chosen studies based on their learning approaches, datasets, and evaluation criteria.
The findings of this review indicate that deep learning and transformer-type models are generally more effective when detecting and preventing cyberbullying because they have improved capabilities of recognizing contextual information and under- standing meaning than traditional machine learning algorithms. Additionally, application of optimization techniques improves performance in identifying bullying by effectively selecting relevant features and tuning hyperparameters.
Although significant progress has been made in our under- standing of the topic, several challenges remain. These challenges include imbalanced datasets, processing of multiple languages and formats and limited ability to explain the rationale for decisions made, as well as ethical issues, and the current lack of real-time implementation of automated systems. Future research should develop integrated explainable, real-time AI- based systems that provide accurate identification of bullying along with proactive measures to prevent it.},
        keywords = {Cyberbullying Detection, Social Media, Artificial Intelligence, Machine Learning, Deep Learning, Transformer Models, Feature Extraction, Optimization Techniques},
        month = {January},
        }