ssrn-5238900

Artificial Intelligence: Transforming Cybersecurity and Combating Cyberbullying Himanshi Dept. of Computer Science &Engineering ChandigarhUniversity Mohali, India er.himanshi.e13362@cumail.in Abstract—Cyber security has turned into a central issue in the digital era. During the last many years, not just the quantity of cyberattacks have expanded essentially, they have likewise become more complex. Thus, planning a cyber- resilient methodology is of vital significance. Traditional security strategies are not satisfactory to forestall information breaks if there should arise an occurrence of cyberattacks. Cyber criminals have figured out how to utilize new procedures and strong tools to hack, assault, and violate information. Luckily, ArtificialIntelligence technologies have been presented into the internet to build brilliant models for shielding frameworks from attacks. Another social abhorrent that can be recognized is bullying, which happens at various phases of society's working and is continually evolving. The fast development of virtual entertainment and online communication platforms has prompted an expansion in cyberbullying, causing significant psychological and emotional harm to victims. Since Artificial Intelligence advances can quickly develop to address complex circumstances, they can be utilized as essential apparatuses in the field of network safety. Artificial Intelligence based methods can give productive and strong cyber defense tools to perceive malware assaults, network interruptions, phishing and spam messages, and information breaks. In addition to these cybersecurity petitions, Artificial Intelligence can likewise be utilized to forestall cyberbullying by observing online interactions, recognizing oppressive language, distinguishing patterns of harmful behavior, analyzing sentiment in real-time and to alarm security incidents when they happen. In this paper, we review the effect of Artificial Intelligence in cybersecurity and prevention of cyberbullying and summarize existing research in terms of benefits of AI in cybersecurity. Keywords—Artificial Intelligence, cybersecurity, cyberbullying, prevention, analyzing sentiment, malwareassaults, phishing. I. INTRODUCTION Artificial Intelligence (AI) plays a significant part in upgrading network protection and battling cyberbullying by utilizing trend setting innovations to identify, forestall, and answer online dangers. Cyber security is not just a problem of IT field. In fact, its scope is very vast.Cyberattacks have grown significantly in both frequency and impact since the first denial-of-service attack in 1988, threatening our increasingly digital world [1].Today everyone is familiar with internet. Smart phones have become indispensable in people's daily lives, even for the illiterate.Cybersecurity involves a range of actions and strategies aimed at protecting cyberspace, devices, software, and data from potential threats.When someone claims that individuals in today's world live online, they are not exaggerating. The internet has developed into an essential component of modern life throughout time. Artificial Intelligence is being used by everyone in their daily lives without sufficient information or awareness. In recent years, the expansion of advanced innovations has prompted expanded cyber threats, going from refined malware attacks to online provocation and cyberbullying. Artificial Intelligence, with its capacity to break down tremendous amount of information quickly and recognize designs, has arisen as an integral asset in network safety and in resolving the issue of cyberbullying. Cybersecurity incorporates a wide range of measures intended to safeguard computer system frameworks, organizations, and information from unapproved access or malicious attacks. Traditional cybersecurity approaches frequently battle to stay up with quickly advancing dangers. Simulated intelligence presents a change in perspective by empowering proactive danger identification and reaction.When cyber security is applied correctly, it can prevent identity theft, data breaches, and other hacker attacks. Therefore, cyber security helps in protecting unauthorized access, modification and deletion of data[7].Artificial intelligence’s core role in cybersecurity is in developing automated incident response systems [3].These systems are equally competent to understand the information, identify potential threats and strive to contain or mitigate the attack in a way that will minimize interference and impact. This is very vitally necessary in case of multiple attacks. Sometimes, people’s involvement often takes too much time for them to respond than the needed time. In web-based conditions, cyberbullying is inescapable, and it alludes to the use of innovation to control, jeopardize, scare, or hurt people. New technologies of artificial intelligence are gradually integrated into identifying cyberbullying based on interactions and contents appeared on the Internet. Additionally, with regards to cyberbullying, simulated intelligence advancements can be utilized to identify hurtful substance, harmful language, or malevolent conduct across different Internetbased stages. Normal Language Processing (NLP) strategies empower Computerbased Intelligence frameworks to grasp setting and feeling, recognizing innocuous correspondence and unsafe way of behaving. Through therefore responding to and solving cases of cyberbullying, artificial intelligence therefore creates better Internet based environments for clients especially vulnerable groups such as children and teens. II. REVIEW OF LITERATURE Due to the advanced level of cyber threats in the modern world, AI has drawn attention from researchers and organizations as a robust tool for cybersecurity improvement. In the last ten years Artificial Intelligence has been established as a key element in shielding against cyber threats including malware and intrusion detection systems. Furthermore, more social problems are being solved with help of AI today, such as the struggle against cyberbullying, in which the possibility of intervention in real time is essential. Aashwi Ranjan Dept. of Computer Science & Engineering Chandigarh University Mohali, India aashwiranjan@gmail.com Manreet Kaur Dept. of Computer Science & Engineering Chandigarh University Mohali, India manreetkaur0711@gmail.com Tanushree Basak Dept. of Computer Science & Engineering Chandigarh University Mohali, India tanushreebasak86@gmail.co m Electronic copy available at: https://ssrn.com/abstract=5238900 -- 1 of 6 -- A. AI in Cybersecurity [1]Artificial Intelligence (AI) has become an essential aspect to strengthen the cybersecurity as the threats have become more sophisticated. Morovat and Panda (2020) reveals on their study based on the AI impact on the cybersecurity with different applications, such as intrusion detection, malware analysis and predictive threat intelligence. AI techniques through ML and deep learning methodologies help systems to detect and mitigate cyberattacks, known threats in addition to real-time patterns. The application of these AI systems becomes highly significant on intrusion detection systems (IDS) which monitor network traffic continually to determine any abnormal behaviour. [3] The usage of AI mechanisms to analyse the high volume of data increases the speed and accuracy of threats identification leading enhancing in overall security process. Artificial Intelligence (AI) has become an essential aspect to strengthen the cybersecurity as the threats have become more sophisticated. Morovat and Panda (2020) reveals on their study based on the AI impact on the cybersecurity with different applications, such as intrusion detection, malware analysis and predictive threat intelligence. AI techniques through ML and deep learning methodologies help systems to detect and mitigate cyberattacks, known threats in addition to real-time patterns. The application of these AI systems becomes highly significant on intrusion detection systems (IDS) which monitor network traffic continually to determine any abnormal behaviour. The usage of AI mechanisms to analyse the high volume of data increases the speed and accuracy of threats identification leading enhancing in overall security process. [6] In her study on AI-based systems for online harassment and cyberbullying prevention, Chaudhary (2024) asserts that Artificial Intelligence (AI) can aid in addressing these challenges. The increasing threat of online harassment, particularly targeting women and marginalised communities, has overwhelmed conventional moderation and prevention mechanisms with the deluge of abusive content on digital media. AI, especially based on Machine Learning (ML) and Natural Language Processing (NLP), can provide scalable and effective approaches for real-time detection and mitigation of harmful behaviours. B. AI in Preventing Cyberbullying [10]Chisholm (2014) presents an overview on the current state of cyberbullying and prevention. With the prevalence of cyberbullying increasing, particularly among youth, it is evident that traditional means for dealing with this issue have been largely ineffective. Many researchers argue that due to the escalation in volume and complexity of problems within online environments, technologies such as artificial intelligence (AI) need to be utilized in detection and prevention of cyberbullying. AI systems can quickly scan through massive amounts of data, such as social media post or profiles, blogs or forums for harmful patterns of behaviour prompt intervention before serious harm can occur. However, AI should be used to compliment human based approaches but not replace them. Challenges with privacy concerns, ethical issues and higher accuracy algorithms need to be furthered explored before AI systems in this area advance. Furthermore, it would improve the paper if case studies or instance of AI application in cyberbullying detection in the real cases could be put into this review. For example, Facebook and Instagram are already using AI device to recognize how offensive the flagged contents are. These kind of instances not only can validate the practicability mentioned by Chisholm but also we can know present situations on AI technology handling with cyberbullying. III. OUTLINE OF ARTIFICIAL INTELLIGENCE Artificial intelligence is the ability of machines to mimic intelligent beings in processing information and solving problems. Artificial intelligence refers to computer systems designed to replicate higher abstractions including reasoning, learning, perceiving, decision making and creative capabilities. Artificial Intelligence is the field aimed at developing machines that are able to solve such problematic situations and make decisions on their own, starting from the speech recognition up to the complicated problem solving and decision making. In its broad sense, AI refers to a set of tools that allow an object or system to perceive its environment, understand it, interact with it, and learn from it. The technological methods used include machine learning (ML) as a technology that helps systems adapt to its data in order to get better, and natural language processing (NLP) as a technology used by the machines to both comprehend human language and generating it as well. AI can be divided into two main categories: Narrow AI (Weak AI): This is a class of AI systems that are built to address one or several functions. These are systems that run on fixed parameters and are far from possessing general intelligence. This ranges from voice assistants such as Siri, facial recognition to spam filters and car black box cameras. General AI (Strong AI): This is the theoretical abstract definition of AI that has abilities that is like the human cognition; which has the capability of performing any of the processes of thought that are available to man. Although, the general AI is one that people strive to achieve, however, until now it has not been realizable. AI is built on the foundation ofML, and is a process in which machines learn about patterns in data and make predictions from such information. Machine learning is a technique where a model is trained with the help of large datasets and then an attempt is made to predict or decide by using that model. Depending on how the model is trained, machine learning can be categorized into three main types: 1) Supervised Learning:In this approach the model is feed with labeled data which means that the input as well as the output is already known. The system adapts to figuring out what input corresponds to the right output and enhancing its work over time. It can be employed when you are working with classification data such as email spam or when working with a regression model on house prices. 2) Unsupervised Learning: In this method, the model is trained with no supervision, in other words the model searches for the latent structure by itself. Customer Electronic copy available at: https://ssrn.com/abstract=5238900 -- 2 of 6 -- segmentation is an example of a clustering algorithm and is a part of the unsupervised learning family. 3) Reinforcement Learning:Here the model takes action on the inputs, generates an output and then is either rewarded or penalized for the action it has made . reinforcement learning is applied in game playing AI and robotics as the system learns best approach through trial and error. AI systems contain behaviours that are inherent following the data they analyze and the algorithms they use. Their behaviour is characterized by: 1) Adaptability: AI structures can correct their performance depending on new inputs. The more data they handle, they enhance their skill of recognizing, classifying, and deciding. 2) Autonomy: Nowadays most AI systems are developed to work on their own, to make own decision and do not require human interferences. For instance, self-driving cars apply AI to build routes, coordinate movements, choose an action depending on the perceptions of the vehicle’s surrounding environment. 3) Precision and Efficiency: AI systems are particularly effective when it comes to repetitive and high volume work, where decisions need to be made based on patterns. In areas such as cyber security it is able to find out suspicious or presumed threats in real time and a lot more efficiently than a human expert. IV. CYBERSECURITY AND CYBERBULLYING DATAFROM INDUSTRY Cyber threats and cyber attacks are slowly becoming common in the world and they are slowly being escalated to even the next level and causing more havoc. From the 2023 Verizon DBIR, one can conclude that both data breaches and cyber incidents are increasing year by year. The frequency of data breaches in organizations across the world was increased to a level of over 4,100 in 2022 not only in the health care, financial, and retail sectors, but also in governmental areas. There were such trends as the shift to the cloud services, the rising density of connected IoT devices and the effects of the COVID-19 pandemic work new patterns of work influencing the frequency of cybercrime. This has been deemed as one of the biggest cyber threats that have been discovered in the recent past half than ransomware. Ransomware is when a person or persons infiltrate an organization in some way or another and then proceed to encrypt that organization’s information and demand that they be paid in order to regain access to the data. The Sophos’s 2023 State of Ransomware Report found that out of the organizations that participated in the survey, 66% of them suffered ransomware attacks in the 2022 financial year, from 37% that were attacked in the 2020 financial year. Many organizations have given in to the demands of hackers but even when the criminals are paid, recovery is not easy and costly and organizations are left with losses. Several high-profile cyberattacks in recent years: 1) Equifax Data Breach (2017): The Equifax breach affected consumers’ personal rights through exposure of names, Social Security numbers, birth dates and addresses of hundreds of millions of customers, 147m to be precise. The attack began from an update that was quite superb, though the update was still awaiting implementation on Apache Struts, a web application framework. Equifax’s problems are major monetary and legal concerns, which do not need to occur with regular patch updates. 2) SolarWinds Attack (2020): The SolarWinds hack, one of the most widely publicized cyberattacks in the past year, was a supply chain attack where hackers compromised the American software-update for the SolarWinds Orion platform, used by many major businesses and governmental organizations. The breach impacted many sensitive departments in the United States federal administrations such as the Department of Defense and Homeland Security. As many as 18,000 customers could have been affected by the malware. This hack exposed weaknesses in global supply chains, showing how third-party software can be used for large- scale spying. It also highlighted how cyber threats are constantly evolving, pushing governments and companies to strengthen their security measures. This attack serves as a warning about the risks of relying on widely used software tools, as they can be exploited in dangerous ways. 3) Colonial Pipeline Ransomware Attack (2021):One vivid example of an actual ransomware attack that reached the world’s largest pipelines of Colonial Pipeline, the company operating as one of the largest fuel pipelines in the United States, led to fuel shortages on the American East Coast. It got a hold of the business for some time, and Colonial Pipeline gave up to the hackers and paid a ransom of $4.4 million. It exposed the cracks in core systems and the magnitude of a cyber holocaust to structural essentials. 4) Log4j Vulnerability (2021): A critical bug in Log4j, a Java-based logging tool that is deployed on millions of servers, was turned into one of the critical and latest security incidents. Known as CVE-2021-44228 or Log4Shell, it was an exploit through which the attackers could gain control of compromised configurations. Because of the fact that Log4j is incorporated into millions of devices and apps across the world this vulnerability was a massive risk and hundreds of organizations had to conduct mass patching. The more an organization depends on functions and processes that are improved by technology and new media, the more potential points of failure are given to the attackers. When looking at the 2023 IBM X-Force Index Threats, the manufacturing and Energy industries seem to be attacked most with 27.7 percentages. This is due to the tendency in utilizing operational technologies (OT) which could not adopt enhanced safety profiles similar to IT domains. Health care, financial services, and education sectors remain other primary customers enabled by digital literacy more than in prior years. The healthcare industry, in particular, experienced a 50% increase in cyberattacks in 2022, with ransomware and information theft being the most pervasive threats. Check Point indicates that healthcare Electronic copy available at: https://ssrn.com/abstract=5238900 -- 3 of 6 -- organizations face an average of 1,463 cyberattacks per week—up 74% from the previous year. Due to the heightened risks which are manifested in a high frequency of sophisticated cyber threats, companies are bolstering their security spending. Another report by Gartner also probated that the global cybersecurity spending would cross $188 billion in 2023 due to the apprehension towards complex threat discernment, cloud security, and IAM. Fig 1: Cyber bullying reports over the years Fig 1: Cyber bullying reports over the years with all these tragic cases are prove of the psychological effects of bullying specifically on youths via cyberspace. Due to such incidences many governments, schools and organizations have put in place laws and policies as well as coming up anti cyber bullying programs to check on such incidents. V. A CASE STUDY In relation to the case study, the paper aims at discussing the concept underlying the development of an experienced- based intelligent system that can help establish smart OSNs without cyberbullying in smart cities. With social interactions and lots of services now moving online due to urbanization, problems such as cyberbullying are much more common, and they negatively impact both mental health and crime rates. To address this issue, the study presents the AICBF-ONS (AI-enabled Cyberbullying-Free Online Social Network) system that is an automated solution that applies different AI techniques to detect and possibly avoid cyberbullying on OSN and maintain safer interactions. The proposed system incorporates several strategies to this end as explained below. First, it cleans the data of this OSN by removing unnecessary characters, noises and stop words in order to guarantee that the data is free and ready for analysis. Following the pre-processing step, the system adopts the Chaotic Salp Swarm Optimization (CSSO) method in gaining the data features relevant to the situation. This step is arguably important as it ensures that the input data is ‘well formatted’ for classification and thus improves on the efficiency of the classification process. In order to classify the log entries, there is used Stacked Autoencoder (SAE) model that is optimized with Mayfly Optimization (MFO) algorithm. This optimization is beneficial to the model to detect the cyberbullying with higher accuracy because of the selection of the parameter. As can be seen in the testing, the AICBF-ONS system was evaluated based on two types of datasets and the results proved that the model achieved high precision, recall, F1 scores, and the results are superior to those of K-CNN and Bi-GRU models. The findings explain why the system outperforms the others in identifying cyberbullying, which is precise in controlling unpleasant exchange in OSNs. Altogether, the AICBF-ONS system delivers the efficient approach to the developing issue of cyberbullying in smart cities’ digital environments. Its ability to select only relevant features and then apply various classification models is a reason it is one of the best approaches to automated detection of cyberbullying. As for the limitations, the study recommends that future research could extend the enhancements made in a large-scale data environment and consider other methods related to isolation of outlying samples as well as clustering, in order to increase scalability and stability levels. VI. ARTIFICIAL INTELLIGENCE TECHNIQUES IN CYBERSECURITY AND PREVENTION OF CYBERBULLYING AI’s role and positive impact on strengthening cybersecurity measures, threats identification, response to cyberattacks, and concept of overall cyber security is beyonda doubt. Some of the most prominent AI technologies used in cybersecurity include: 1) Machine Learning (ML) for Threat Detection:: Machine learning is one of the most popular applications of ‘Artificial Intelligence’ within the cybersecurity field. Machine learning algorithms process large amounts of historical data to recognize the spectrum of ordinary and suspicious network, system and user activity. This enables the trained ML models to identify new threats that include malware, ransomware and phishing. 2) Artificial Intelligence IDS: The IDS involves the use of advanced technology to monitor the network traffic in case of unauthorized access or an attack. AI-enhanced IDS employs the use of anomaly detection which involves a machine learning algorithm that is designed to identify ‘abnormality’. New information integrated into the networks automatically help the AI-IDS to learn about the intrusions and when there is suspicious activity, raise the alarm. 3) Natural Language Processing (NLP) for Phishing Detection: Phishing is one of the most prevalent cyber threats in which a user is tricked into providing personal details. The AI is being applied to discern phishing emails and websites by using the Natural Language Processing (NLP) to dissect the words and general format of messages. Machine learning can be trained to find markers in spam mails and emails that include and contain phishing intention such as language and formatting styles of the message, and Electronic copy available at: https://ssrn.com/abstract=5238900 -- 4 of 6 -- the URLS present in the message so that people and organizations will be warned of scams before they succeed. 4) Insider threat: Whereas insiders are employees and contractors with unauthorized access to their organizations for malicious intent, behavioral analysis for their detection is almost impossible with conventional cybersecurity solutions. Behavioural analytics powered by artificial intelligence constantly run checkups on users for any signs of suspicious behaviours. Similarly, to achieve the goal of early detection of insider threats, AI creates behavioral profiles of users and looks for deviations that might indicate that a user is not only knowingly sharing sensitive information with competitors but also using company’s login details at prohibited times. 5) Automated Incident Response:AI is also being included in automated incident response systems. Such systems employ artificial intelligence in identifying threats and take measures in responding to them in real time. For example, AI based solution for protection of a network can identify a break in by a hacker or virus and shut the invasion down, or isolate tainted devices or quarantine viruses before they spread. AI Technologies in the Prevention of Cyberbullying: Apart from using it in the technical aspect of cybersecurity, AI is more and more incorporated to solve social issues like cyberbullying which has psychological impacts on users. Due to the increased amount of interactions and anonymity that Internet allows, manual quality control of abusive content is practically impossible. AI can provide large-scale and automated means for tackling the negative activities performed by users in their interactions. 1) Natural Language Processing (NLP) for Detecting Abusive Language: The primary technological advancement that is applied and implemented in the case of cyberbullying is the Natural Language Processing technique. The use of NLP allows for machines to see, comprehend, and analyze text data specifically what is considered abusive, threatening or harassing language in cyberspace [5]. Sentiment and intent analysis can be trained, using NLP models, and the identified text, so that AI systems can filter out potentially dangerous content for the attention of moderators or delete them directly. 2) Sentiment Analysis for Early Detection:The general meaning of a text is not sufficient in sentiment analysis as it also includes the identification of the emotional intended message behind the text [7]. AI systems are now capable of observing the content of the ongoing online discourse and identifying specific negative affects, such as anger, sadness, hostility in such a discourse, which are potential predictors of cyberbullying. These are the early signs that if detected, AI can prevent bullying from escalating before it happens. 3) Deep Learning for Contextual Understanding:Whereas initial anti-cyber bullying system implemented simple keyword filtering techniques, the modern deep learning algorithms take into consideration the actual conversation context. Cyberbullying can include the use of sarcasm, coded language, and indirect insult, all forms that are difficult for conventional rule based systems to pick up. We see that deep learning models, which are built with a structure similar to the human brain, help in perceiving the context of language better. 4) AI-based Monitoring Tools:Most social media platforms now employ AI based monitoring tools for the identification of cyberbullying. These tools pick posts, comments and messages in user generated content for any negative behaviour and are capable of reporting or deleting such content on their own. For instance, Facebook and Instagram have integrated machine learning and artificial intelligence algorithms to screen and moderate user relations so that human moderators can handle complaints about bullying incidents more efficiently than automated bots. These AI-based monitoring systems do not only serve as means to safe guard the users and prevent them from falling prey to such scams but also feed information back into the AI system, under which they can become increasingly more efficient and effective with time as the data collected reveals people’s behaviors more and more accurately. 5) Real-time Interventions and User Support:Real Time Solutions and Users Support Moreover there is application of AI for real time interventions against cyberbullying. Using configured sets, AI systems can even notify the users when they are about to publish a toxic message to the target personality, allowing the users to change their mind. Such systems are usually implemented to ensure that users of the sites engage in typical acceptable behaviors which are not deemed violent by other users. Fig 2: AI-based Cyber security solutions AI technologies are proving to be effective in cybersecurity and the prevention of cyberbullying as the technology offers efficient, accurate, and reliable ways of identifying threats. In cybersecurity, applications based on artificial intelligence help improve threat identification, as well as minimize human participation in responding to threats, and detect patterns in incidents that might otherwise remain unseen. While in the context of social networking and presence on the Internet, AI based systems identify toxic behaviour, monitor and prevent bullying, and offer immediate assistance. Electronic copy available at: https://ssrn.com/abstract=5238900 -- 5 of 6 -- VII. CONCLUSION Encrypting Artificial Intelligence (AI) in cybersecurity and combating cyberbullying offer feasible approaches to two premier issues of digital era. In the face of increasingly complex and sophisticated cyber threats, AI-based security solutions provide organizations with a means to identify and manage security risks, as well as develop means to counteract security threats at an extraordinary rate and with high precision. From surveillance of crucial infrastructures and prevention of cyber-threats to identification of aggressive narratives on social media-AI is continuing to become a major force in building a protected cyber-space. I have found out that real time analysis of data and ability to learn from it is one of the biggest advantages of AI implementation in cybersecurity. With advancement in the sophistication of the technology used in machine learning algorithms, the above areas like malware, phishing, and network security will record even better results. In the same way, the available Artificial Intelligence moderated tools on social media sites are helping in necessary moderation for cases of cyberbullying. Such systems can indicate cases of toxic behaviour before they become worse and might help to lessen the amount of emotional abuse given to people. In addition, AI-driven tools are also being used to detect fake accounts, prevent the spread of disinformation, and enhance overall user safety across online platforms. As AI technology continues to improve, its role in protecting digital spaces from cyber threats and harmful behavior will only expand, offering more comprehensive solutions to the growing challenges in cybersecurity. These innovations not only ensure a safer online environment but also help foster healthier interactions and greater trust among users. That being said, there are still some issues that need to be solved, like how to deal with AI’s ability to recognize different forms of cyber abuse, new forms of security threats, and so on; However, future holds great promise in terms of using AI in these fields. As future advancements are made, AI will advance it its ability to be better utilized, fair and accurate in its performance. Businesses, governments and social media are already embracing and fund AI solutions, and as these technologies become more advanced, stronger more proactive measures are already being taken to secure enterprises from hacking and to safeguard the individual from cyberbullying. Moreover, the IA and specifically the developed solutions prevent cyber threats and cyber bullying but with the aim to create the safe world for the further generations. With the advancement of the technologies AI provides the hope in the future when technology will not only protect from the cyber threats, it will help create a better and safer world for people online. By one hand, AI will without doubt enhance its role as one of the major building blocks for digital safety and anti cyber threats, through inter industry and governments’ cooperation with technological experts. REFERENCES [1] K. Morovat and B. Panda, "A Survey of Artificial Intelligence in Cybersecurity," 2020 International Conference on Computational Science and Computational Intelligence (CSCI), Las Vegas, NV, USA, 2020, pp. 109-115, doi: 10.1109/CSCI51800.2020.00026. [2] Erokhina, E. V., and T. V. Letuta. "Juvenile Cybersecurity and Artificial Intelligence System." 2nd International Scientific and Practical Conference on Digital Economy (ISCDE 2020). Atlantis Press, 2020. [3] Vyawahare, Harsha, Sarika Khandelwal, and Seema Rathod. "Artificial Intelligence in Detecting and Preventing Online Harassment." AI Tools and Applications for Women’s Safety. IGI Global, 2024. 14- 35. [4] Walli, S. A., Kang, B.-G., & Nam, Y. (2024). Innovative Artificial Intelligence Solution as Game Changer in Cyberbullying Detection and Prevention. Artificial Intelligence in Cybersecurity, 1, 52-59. [5] Sarker, Iqbal H., Md Hasan Furhad, and Raza Nowrozy. "Ai-driven cybersecurity: an overview, security intelligence modeling and research directions." SN Computer Science 2.3 (2021): 173. [6] Chaudhary, Anita. "AI-Based Online Harassment and Cyber Bullying Prevention System." Impact of AI on Advancing Women's Safety. IGI Global, 2024. 24-32. [7] Milosevic, Tijana, et al. "Effectiveness of Artificial Intelligence–Based Cyberbullying Interventions From Youth Perspective." Social Media+ Society 9.1 (2023): 20563051221147325. [8] Azeez, Nureni Ayofe, et al. "Cyberbullying detection in social networks: Artificial intelligence approach." Journal of Cyber Security and Mobility (2021): 745-774. [9] Perera, Andrea, and Pumudu Fernando. "Accurate cyberbullying detection and prevention on social media." Procedia Computer Science 181 (2021): 605- 611. [10] Chisholm, June F. "Review of the status of cyberbullying and cyberbullying prevention." Journal of information systems education 25.1 (2014): 77. [11] Walli, Salma A., Byeong-Gwon Kang, and Yunyoung Nam. "Innovative Artificial Intelligence Solution as Game Changer in Cyberbullying Detection and Prevention." Artificial Intelligence in Cybersecurity 1 (2024): 52-59. [12] Reddy, L. Rajeshwar, et al. "Detection of Cyberbullying Across Social Networks." 2024 International Conference on Intelligent Systems for Cybersecurity (ISCS). IEEE, 2024. [13] Shankar, Karthiga, et al. "Cyberbullying Detection in Social Media Using Supervised ML and NLP Techniques." Communication and Intelligent Systems: Proceedings of ICCIS 2021. Singapore: Springer Nature Singapore, 2022. 817-828. [14] Ravikumar, S., et al. "Detecting cyber bullying using Ml algorithm." 13th International Conference on Advances in Computing, Control, and Telecommunication Technologies, ACT 2022. 2022. [15] Battula, Sai Poojitha. Cyberbullying (Hate Speech and Offensive Language) detection using machine learning. Diss. CALIFORNIA STATE UNIVERSITY, NORTHRIDGE, 2024. [16] Kapoor, Sumit Kumar, et al. "Security and Threat in Online Social Networking." Online Social Networks in Business Frameworks (2024): 449-477 [17] Liaqat, Muhammad Saeed, et al. "Exploring Phishing Attacks in the AI Age: A Comprehensive Literature Review." Journal of Computing & Biomedical Informatics 7.02 (2024). [18] Al-Marghilani, A. Artificial Intelligence-Enabled Cyberbullying-Free Online Social Networks in Smart Cities. Int J Comput Intell Syst 15, 9 (2022). Electronic copy available at: https://ssrn.com/abstract=5238900 -- 6 of 6 --