The role of artificial intelligence in radicalisation, recruitment and terrorist propaganda: deconstructing violent extremism and reimagining counterterrorism in contemporary digital ecosystems

Abstract

Contemporary digital ecosystems present a paradigmatic shift in how extremist organisations operationalise artificial intelligence (AI) technologies to facilitate radicalisation, recruitment, and propaganda dissemination. While AI has catalysed transformative advancement across multiple sectors, its dual-use characteristics render it susceptible to weaponisation by violent non-state actors. Extremist groups systematically exploit AI-driven recommendation algorithms, behavioural profiling mechanisms, and generative content systems to identify and target psychologically vulnerable populations, thereby circumventing traditional counterterrorism methodologies. This study examines the multifaceted mechanisms through which AI technologies encompassing machine learning, natural language processing (NLP), facial recognition, and synthetic media generation facilitate the radicalisation trajectory outlined in contemporary theoretical models of violent extremism. This qualitative exploratory research employs secondary data collection methodologies integrating academic literature, cybersecurity reports, government publications, and open-source intelligence (OSINT) from publicly accessible digital platforms. The analytical framework synthesises content analysis and network analysis to systematically identify extremist rhetorical strategies and information dissemination pathways. Psychological theories of radicalisation, coupled with algorithmic media theory and social network analysis, inform the examination of how AI technologies exploit cognitive vulnerabilities and amplify ideological messaging across transnational digital networks. Temporal scope encompasses the period from the early 2010s to the present, with particular emphasis on developments within the preceding five years. The investigation reveals that AI-enabled technologies facilitate radicalisation through algorithmic amplification of emotionally provocative content, behavioural analytics enabling precision targeting of at-risk individuals, and generative systems producing synthetic media (deepfakes, AI-crafted audio) that circumvent content moderation mechanisms. Case studies of ISIS’s algorithmic recruitment strategies, Al-Qaeda’s generative AI-powered propaganda, Taliban’s encrypted-platform messaging, and far-right extremist memetic warfare demonstrate systematic weaponisation of AI technologies. Evidence from geopolitically significant regions, particularly Kashmir, Afghanistan, Syria, and Western democracies, illustrates how extremist organisations leverage AI to create personalised recruitment campaigns, establish echo chambers, and manufacture synthetic narrative content while maintaining operational security. The analysis identifies critical gaps between conventional counterterrorism approaches and the velocity of AI-driven radicalisation. The accelerating sophistication of AI-facilitated extremism necessitates comprehensive international cooperation frameworks, ethics-oriented regulatory architecture, and AI-powered countermeasures. Current legal structures remain fundamentally inadequate to address borderless, rapidly evolving threats. The paper advocates for integrated multilayered responses, including AI-enabled early detection systems, transparent algorithmic governance, international intelligence-sharing mechanisms, and digital literacy initiatives. Critical ethical considerations, surveillance implications, privacy protections, freedom of expression, and algorithmic bias must be balanced against security imperatives. Successful mitigation requires coordinated efforts among governments, technology platforms, international organisations, and civil society to develop ethically designed AI tools for counter-extremism, while establishing robust accountability mechanisms that safeguard human rights in the digital age.

Introduction: artificial intelligence and emerging frontiers in radicalisation, navigating challenges and exploring prospects for counterterrorism

Over the past few decades, Artificial Intelligence (AI) has emerged as a transformative technology driving rapid innovation and advancement across multiple industries. AI, defined as computational systems capable of simulating human cognitive functions, has become integral to diverse sectors ranging from healthcare and finance to education and public administration. The widespread adoption of AI technologies, including machine learning, deep learning, natural language processing, and computer vision, has generated significant breakthroughs in medical diagnosis, financial prediction, personalised learning, and automated decision-making systems. AI is getting more deeply embedded into everyday tools and services such as in healthcare and finance, education and entertainment to make people more productive, more efficient and more convenient (Klingberg, 2022). With the help of AI-driven technologies, Machine learning (ML), deep learning, natural language processing (NLP), computer vision, robotics, etc., have resulted in outstanding advancements that did make breakthroughs possible like diagnosing in medical, predictions in finance, personalised learning, automated customer service, etc. Take healthcare, for example; AI is being leveraged in diagnostic tools to interpret medical imaging, predict patient outcomes or even develop customised treatment plans for each patient. When it comes to finance, using AI for the purpose of fraud detection, algorithmic trading and risk assessment has become critical (Graziani, 2023). Thanks to them, AI teaches our children, provides automated tutoring and simplifies administrative tasks in education. With the integration of AI, it has reached considerable improvement in efficiency, cost efficiency and accuracy, and has been widely used as a transformative technology in multiple industries (Da Silva et al., 2023). However, AI is a potential dual-use technology—its wide use has increased in every sector, whereas its misuse creates a lot of threats. With these technologies becoming more prevalent, so has the ability of them to be misused. Amongst the most high-pressure issues in this regard is the use of AI by extremist groups to further their ends. There are many things with which AI is capable of, one being processing lots of data, personalising content, and automated communication, which makes AI a very powerful tool against radicalisation as extremists do not have to waste time or resources trying to reach an individual; instead, they are able to, yes, use the same amount of data from the individual to provide them with customised content and automate the radicalisation process by doing so (Casebeer et al., 2018). The ethical considerations of AI, in particular, are a dual-use technology, particularly in regard to national security and public safety. The good that AI can do for us is also limited and can be used for evil, to manipulate information, target vulnerable persons and trick some people with convincing propaganda, among other threats. Using AI in such a way amounts to an increasing danger to its ability to spread far-right extremist ideologies, create echo chambers, and magnify cries of radicalisation over digital media (Zhaksylyk et al., 2021).

The AI-extremism crisis: understanding digital radicalisation in the 21st century

With advancing artificial intelligence technology being adopted at various stages of the digital ecosystem, extremist groups are increasingly employing the technology to radicalise, proselytise, and spread their ideologies. With the rise of social media along with social media algorithms that integrate AI, they have become a rich ground for extremism to spread. The use of AI tools (like recommendation engines, content generation algorithms, and automated bots) allows extremist groups to reach huge, expanded audiences, manipulate human emotions through will and force people in ways that would not have been foreseen before. Another concerning element of using AI in radicalisation is its capability to recognise and take advantage of psychological vulnerabilities (Calvert, 2024). Through analysing huge amounts of personal data, the AI algorithms can tailor messages and content which would rather appeal to this particular person’s emotions, beliefs and grievances. An AI system can, for example, detect people who become up, run down and/or angry, which are emotions that extremist groups can take advantage of and recruit new members or intensify the existing followers’ dedication to violent ideologies. In addition, sophisticated propaganda, such as deepfake videos, fake news, and automated social media accounts have been possible to be created through AI technologies (Mølmen and Ravndal, 2023). AI-generated tools are used to spread extremist messages, build ripple of thoughtless narratives, and ultimately destroy the public’s perception. Social media algorithms enhance the reach and impact of such content as the content which generates the highest levels of engagement is prioritised irrespective of its accuracy or possible harm that it can cause to society. The implications of an AI enabled radicalisation process are security critical. It makes the spread of extremist ideologies go faster, and battle receding from hunting them will become harder to win. The methods of counterterrorism using traditional AI have now become less valid because of the AI’s capability to automate the spreading of extremist content. In addition, because digital platforms keep you anonymous, authorities are unable to find where to track down and intervene on an AI driven radicalisation campaign (Little et al., 2021; Wirsing, 2004). Finally, the way that AI has been used perversely globally for radicalisation is one that has national security, governance and ethical implications. Protecting the public from the harmful use of AI is a big task and there is a constant challenge in regulating AI for governments and security agencies (Raitanen and Oksanen, 2019).

From Kashmir to the world: understanding transnational AI-enabled extremism

The purpose of this study is to examine the usage of Artificial Intelligence (AI) in logics, recruitment and propaganda prose in the context of extremist groups in a global perspective. Nevertheless, the analysis of the article as a whole evaluates such phenomena across the world, and focuses on areas where radicalisation processes do occur online (Pandya, 2020). In particular, these regions have experienced significant AI-driven extremism incidences, within a large part of which the digital ecosystem is central to promoting extremist ideologies. At the same time, online radicalisation has had high-profile cases in European countries, such as ISIS and other jihadist organisations, still utilising AI in order to recruit individuals from various backgrounds. At the same time, the rise of right-wing extremism in Europe is an equally major challenge, which AI allows to personalise the delivery of extremist content to disenfranchised individuals (Legide, 2022). The spread of the radical messages of extremist groups has been occurring through social media and AI tools in the United States and Canada (North America). This includes the recruitment of people into violent right-wing and left-wing extremist thought and the use of AI to control election outcomes through the use of manipulated public opinion. The digital ecosystem in North America, with its diverse and engaged online communities, offers a rich environment for AI-driven radicalisation. In the Middle East: Because the Middle East has remained a central area for AI-driven radicalisation, extremist groups such as ISIS, Al Qaeda, and their home groups have used AI technologies on a large scale to spread their narratives and to recruit people regionally and globally. The reason for the key significance of this region of study is geopolitical destabilisation and the existence of online extremism in this region. The study sheds light on how cultural, social and political contexts impact the way in which extremist actors utilise AI in extremist activities by highlighting these regions. The evolution of AI’s role in radicalisation is addressed in the study, from the early 2010s until the present day, outlining the process of AI’s adoption into extremist online strategies. This is the time in which AI technologies are developing very rapidly and also becoming increasingly relevant to digital space dynamics, specifically with respect to radicalisation and extremist recruitment (Khaire, 2023). Early 2010s: This was the beginning of AI and automated tools adoption by extremist groups. While still in a nascent state, AI would have served as a platform for groups to spread their ideology, like those in the group ISIS, which started promoting their ideology on social media platforms. Early uses of AI included basic content curation and basic targeting (Sofi and Dogra, 2019). In the Mid to Late 2010s, the applications of AI technologies to radicalisation grew. In the mid-2010s, however, sophisticated AI-driven algorithms started to allow extremist groups to arrive at larger, more specific audiences. Extreme and extremist propaganda took on bigger proportions of scale and reach by means of more prevalent AI-based recommendation systems, deepfake videos and automated bots. The latest developments in natural language processing, machine learning and content automation have given groups the ability to make propaganda more personalised to people, control emotions and recruit potential recruits more effectively (Balcerowicz and Kuszewska, 2022). Finally, there has been a drive to integrate AI into digital security systems and counterterrorism efforts. However, using AI against extremist actors requires much more work to mitigate the negative abilities of AI misuse. This study limits itself to this temporal scope for the sole reason that it emphasises the speed at which technology is evolving and the concomitant rise in extremist AI usage (Neelamalar and Vivakaran, 2019). Specific attention is paid to developments in AI over the last 5 years, as the previous 5 years have been both the most successful and the most risky for AI capabilities and associated security risks, respectively.

Multilayered threats: analyzing AI-generated propaganda, recruitment, and radicalisation content

The main scope of the content of this study deals with the role played by AI in the Web promotion of radical content and its contribution to extremist recruitment (Fernandez et al., 2019). This covers the analysis of how the use of AI interferes with digital platforms for the spread of extremist ideologies and how AI technologies are utilised to channel individuals to violent extremist groups. It is centred on AI and the role it plays in maximising the reach (maximising virality) of extremist content. This involves the usage of social media algorithms to broadcast important messages, the creation of deepfake movies or phoney news tales, and the application of computer-generated content (Herath and Whittaker, 2023). The research will discuss how these technologies give us access to radical content, which we can become persuaded even further by and become more immersed in, leading to the further spread of extremist ideologies. Recruitment through AI: The study will investigate how AI algorithms are employed in order to recruit or target people who are susceptible to radicalisation. This includes behavioural data analytics so that extremists can draw a personal profile of individual users based on social media trails, search trails and online interactions and create personalised recruitment campaigns. I will analyse the effectiveness of AI in forming emotional ties with hopeful recruits and incenting them to join extremist groups.

Research questions of the study

The current study, still in response to the accumulating pressures of the application of artificial intelligence (AI) on the radicalisation, recruitment and propaganda processes of extremist groups, is guided by four key research questions. The most important issue is how AI may play a role in the process of radicalisation itself and how the mechanisms through which AI-enhanced technologies (personalisation algorithms, high-quality data mining, and automated content creation) can be used to target and identify vulnerable individuals should be explained. This research question will address the extent to which psychological and behavioural profiling techniques that are implemented in AI systems personalise and amplify radicalising messages, thereby amplifying extremist ideologies. The second research question changes to the nature of AI as the part of the recruitment and propaganda efforts of the extremist organisations, and, specifically, the effectiveness and the use of AI-based tools such as chatbots, synthetic media (deepfakes), and autonomous social media bot accounts are offered as a means of delivering the messages of the extremist organisations and enhancing their persuasiveness effect, at least in the context of the digital ecosystems in which the radical discourse is widely common. The paper subsequently challenges the extent of security, ethical and legal implications the contribution of AI to radicalisation may impose, challenging both the strain on national security apparatus, individual rights and primarily the rights to privacy and freedom of speech, and the challenges of balancing regulatory constraints and the safeguarding of the fundamental civil liberties. Finally, the study also attempts to suggest and critically analyse a collection of countermeasures that can address the facilitative impact of AI in extremist actions, which includes AI-based detection algorithms, systems that facilitate international and global cooperation, ethical standards of AI application in countermeasures, and measures to enhance public digital literacy and resilience towards radical content. The study is expected to construct a complex image of the multifarious effects of AI on contemporary extremism and chart the means of successful and ethically possible counterterrorism strategies with the help of these interdependent research questions.

Research objectives of the study

• To investigate the idea of artificial intelligence technologies and their application to the radicalisation of vulnerable individuals, one should consider such aspects as personalisation algorithms, profile-based behaviour forecasting, and the automated creation of the content used by extremist groups.

• To evaluate the significance and effectiveness of AI-based technologies, including chatbots, deepfakes, and automated social media pages, in relation to the role of extremist recruiting or propagating propaganda on the internet.

• To critically assess the national security, ethical, and legal concerns of AI’s role in radicalisation, with a focus on its impact on privacy, freedom of speech, surveillance, and regulatory governance systems.

• To prescribe and evaluate comprehensive countermeasures to this, such as AI-enabled detection systems, global interoperability, and the moral values of applying AI in counterterrorism, as well as community digital literacy initiatives to limit extremists’ use of AI.

Theoretical framework: integrating AI, radicalisation, and counterterrorism

In this analysis, a comprehensive theoretical framework is utilised that combines both traditional and modern theories in the analysis of the complex role of artificial intelligence (AI) in the radicalisation, recruitment, and propaganda of extremist groups (Omotoyinbo, 2014). The Staircase to Terrorism by Moghaddam is the core of the framework in which radicalisation is a psychological process that can be gradual and can be examined considering the use of AI-powered personalisation and behavioural profiling to drive people through extremism (Fernandez et al., 2018). In addition to this psychological prism, the Social Network Theory offers the required information regarding the AI-based network effects, which implies how the amplification of algorithms, the utilisation of bots and the recommendation system redefine the relationships between the flows of information, form an echo chamber, and unite extremist groups on the internet (Raja et al., 2025). Furthermore, the Algorithmic Media Theory states the influence of AI on the consumption habits of the media through giving preference to the content in relation to engagement and producing a bigger exposure to extreme ideas (Tanoli et al., 2022). In combination with Ethical and Legal Governance Theories, the research has an opportunity to wrestle with the normative question, such as the protection of rights, accountability regimes, and the system of global regulation in terms of AI application in counterterrorism (Neo, 2022). An intricate part of these theories is an interdisciplinary nucleus that allows introducing the idea of AI as the conditioner of radicalisation and the mechanism of its control into the context of the current security issues with a certain degree of sophistication (Mirchandani, 2018).

Literature review: AI and its role in online manipulation

As stated by Idahosa (2017), the present extremist groups also use and deploy (AI) artificial technology in using the same to reboot the online behaviour to further their set of end agendas. Machine learning, natural language processing (NLP), facial recognition and automated content building were used, which could be further utilised and would help in personalising the way in which extremist messaging would be carried out (Williams and Tzani, 2024). For example, by automating the process of analysing data through machine learning in particular, as well as making it possible to go through tremendous datasets searching for patterns in people’s behaviour that can be used to seek out vulnerable people that extremist groups can use (Zhaksylyk et al., 2021). This is often done by these algorithms using social media, search behaviour and any other interactions online and curation of the content that chases a person’s vulnerabilities and leaves them more vulnerable to radicalisation. Extremist organisations use natural language processing (NLP) technologies that are trying to generate and understand human language to create catchy messages that trigger the emotions of the audience (Putra, 2017). For instance, the people whom the automated messaging systems and AI chatbots powered by AI will sail for the purpose of simulating human likeliness to build feelings of belonging and trust so they can feel ‘like the rest of humanity’ and help them to enrol members for radicalisation more quickly. Further, NLP tools are also deployed to check social media posts for frustrated or irritated users of a particular set of grievances or ideological leanings as extremist and, thus aiding in the passage of targeted propaganda (Idahosa, 2017). Facial recognition technology, for example, enables extremist organisations to implement another form of targeted recruitment and surveillance, in which other critical tools are also adopted by extremist organisations, including differential casing or analogies of case, ROI, or rate of interest. It can be applied to monitoring public spaces to detect individuals who present themselves as potential targets to be radicalised and also present themselves to be in accordance with a particular extremist cause online (Montasari et al., 2022). The propaganda, on the other hand, is also composed using the systems of a very broad spread of the automated content creation. However, these systems make it possible to produce large amounts of content such as text, pictures or video with little or no people involved. Also, because of this ability, extremist groups have the capacity to create personalised content and scale. They can succeed in being more efficient and effective than traditional means in disseminating their radical ideas. According to the literature, like the cases of ISIS and Al Qaeda, extremist groups have already adopted the use of AI to enhance the refinement of delivery of their content, hence expanding the clients (reach) and making a greater impact than ever. The literature used examples of targeted advertising and deepfake technology to entice online behaviour through AI technologies (Butler and Montasari, 2023). Extremist groups can use targeted advertising with AI-driven algorithms to push messages that are purpose-built for specific individuals based on demographic, psychographic and other data points that they can publish using that algorithm (Sirgy et al., 2018). These algorithms can scan and analyse what users put out and what they like and so are much more useful in persuading the beliefs to the ‘viewpoints’ to suit their worldview. Assuming AI-generated videos that change people’s facial expressions or recite their voice as their own to create fake narratives, advocate propaganda and trick viewers into believing in the artificial has become a complex scenario in their case (Jääskeläinen and Huhtinen, 2020). The most powerful thing about the weapon emanates from extremists who are able to create artificial visual and audio content that make gullible people believe these messages are from real human beings. Additionally, these groups will then exploit recommendation algorithms available on YouTube Facebook or other such platforms to make that content appear as highly visible as possible (Schnader, 2020).

From algorithm to allegiance: understanding AI-mediated radicalisation processes

Radicalisation is a multi-stage process of adopting belief changes and attitude changes, which usually end up with the assumption of extremely violent and radical ideologies. According to psychological theories of radicalisation, radicalised individuals link their identities and grievances to those led by social networks that radicalise and perceive themselves to be mistreated, which in turn is enhanced and networked through social media (Aprin et al., 2021). To this end, AI increases the size and focus of the scope of communities that extremist groups can target for individuals who would likely be radicalised. Based on algorithms, AI is able to notice who has some behaviour, attitude, and even some interests that one can infer the predisposition to extremist ideologies (Williamson et al., 2019). AI systems equipped with behavioural analytics can observe online behaviour patterns and predict which people are more likely to be drawn into recruitment. For example, the way people socialise socially on social media, search and browse history are analysed to determine which people are likely to surf extremist content (Blasiak et al., 2021). Extremist organisations have the capability to predict who will become extremists, therefore, they can guide their recruitment toward those people believed to be most trusting to their message. Therefore, algorithmic targeting is one such powerful tool where AI acts so as to shift just the right personalised content to the right people based on their own interests, emotions, psychological state and likes. With the ability to mine data about the interactions online, AI catalogues and modifies content and tweaks its content to be about what individuals believe today and eventually alters their mental state to more extreme believes (Hossain, 2015). The introduction of radical ideologies through filter bubbles (that are algorithmic; they suggest content matching an individual’s previously proven decision making, i.e., their underlying activity) makes it harder for the people to find the counter-narratives. Another attribute of AI is creating echo chambers, which are digital places where content matches the ideas, and the opposing ideas are bypassed (Proctor, 2021). The extremist ideologies, however, grow strongly as they are magnified inside echo chambers as belief in them becomes socially reinforced to the people. Studies show that AI keeps selecting the same content that delivers extremist messages (Trehan, 2002). With time, the radical positions the content carries, in fact, promote extremism among people who view them. Facebook, Twitter and Telegram are widely frequented by these groups for spreading their message and, in some cases are being used as crucial platforms to recruit and propagate messages from extremist groups. Case studies that involve the exploitation of AI technologies by ISIS and Al Qaeda, so individuals are recruited. For example, a video-focused strategy of ISIS capitalising on the power of online recruitment, targeted issues in online forums, and encrypted communications with the help of AI-enabled tools made it possible to operate globally (Proctor, 2021). They rely on artificial intelligence (AI) based algorithms to scour audiences to find these marginalised people with grievances and who are hurting socially or economically in the hope of finding potential recruits. The influences of Propaganda and AI in the leading of Ideologies. One of the problems of employing AI to support propaganda is that it has made the use of such propaganda much more efficient, personal and ultimately more effective compared to conventional radical messaging (Conway, 2017). The first advantage is one of the unique features of AI that can generate content based on the special psychological and emotional bytes and needs of the individuals. Instead of generic messaging, AI can analyse an enormous amount of personal data to make propaganda that speaks directly to an individual’s fears, desires and grievances. AI is also good at generating content assistance, particularly when we are talking about content, material like articles, videos, memes, and social media posts, or as much content as possible and as fast as possible. The capability to propagate their messages as extremist groups on the internet without having to do so in real life and direct their messages to the millions of people on the internet without having to do the same in mass meetings (Thomas-Evans, 2022). However, out of all the upshots in this field, the most noticeable one is the deepfakes, that is, AI-generated videos that fabricate or falsely manipulate reality into realistic footage. If they are given the wrong videos, they can be used to tell false narratives, and to mislead the audience to build a false version of reality that benefits the promotion of extremist ideologies. Also, AI is involved in the construction of fake news (Biersteker, 2007). The negative impacts of AI-generated content implication on rapid dissemination of misinformation, public trust and social cohesion have been made. The extremist groups with imitation goals are using AI to make up stories that exploit existing social, political or economic tensions and deepen the polarisation of public opinion (Bamsey and Montasari, 2023). The second form of other AI-driven tactics involves using bots to amplify extremist content. These are automated accounts using machine learning algorithms to propagate propaganda, support certain messages and interact with the users as one human being. The more extremism is seen as popular, the more mainstream it seems. Extensive literature supports the conclusion that the instrumental value of AI in spreading propaganda through visual, textual, and auditory type content is extremely high. As such, in this age of the digitised globe, content is consumed passively and in haste (Casebeer et al., 2018). AI will not replace the need to make large human resources and efforts that traditional propaganda requires; however, it will accelerate and amplify the propaganda campaigns without necessarily making them more detailable; they might even become more pervasive.

Security paradoxes: how AI countermeasures create new vulnerabilities

Increasingly, AI is being used for propaganda and radicalisation that are morally and security-wise of grave concern. This debate topic revolves around the ethical issues with the changing emotions, producing of deep fakes, and the impact on freedom of speech (Pfeiffer, 2012). In this case, questions arise about what, outside the digital realm, is acceptable or impossible behaviour for behaviour AI can modify a person’s emotions and beliefs based on the content that it can manipulate. Though the use of AI led to the creation of clever deepfake videos, fake news stories, and so on, there remain the most outstanding challenges on the front of information integrity and the trust we need to put in digital media (Wilner and Dubouloz, 2011). This problem deals with how the right to free speech can coexist with the duty in order to reduce the adverse effects of such extremist ideologies. We should hold on to our right to free speech, but that means attending to the lines that we know to be morally, ethically, and politically insupportable; eventually, such AI-generated propaganda will be able to manipulate people into assuming extreme positions and, by extension, bringing about violent outcomes, and for these governments and the companies creating the tech need to put in place frameworks for dealing with those ethical issues. Firstly, various security concerns are involved in using AI in radicalisation. Public safety is threatened by AI as it can be used in surveillance, target an individual and spread extremist content (Wirsing, 2004). The same technology is now being used by the government to build AI-powered surveillance systems that monitor potential threats. Still, the extremists have begun using it to monitor activists, journalists, or anyone else they wish. Additionally, mass surveillance powered by AI is of concern with regard to the privacy of citizens, and it violates their rights to civil liberties and human rights. Moreover, under the sub-councils, there is a clear way for the use of AI in the protection of national and global security (Maskaliūnaitė, 2015). AI being utilised for the purpose of influencing manipulating public opinion, using elections for their benefit and raising extremism is a risk to the social stability and deteriorates the trust in democracy and encourages a social uprising. Further, the more AI is employed in securing the cyber world, the more the adversaries can employ AI vulnerabilities against their targets to perform fine-grained attacks (Clubb, 2019). However, extremists’ use of AI in radicalisation, recruitment and propaganda is a complex and large challenge. However, AI can also be an invaluable tool for malefactors that, thanks to the bad misuse of the technology, give rise to ethical and security issues, which can be solved by resorting to properly designed law regulations, development of the technology itself, and international cooperation.

Methodology

This research uses a qualitative, exploratory method of research to investigate the use of Artificial Intelligence (AI) in the processes of radicalisation, recruitment, and propaganda activities of extremist organisations. For example, the qualitative method is used in the exploration of complex, under-investigated issues where in-depth insights into how AI technologies may contribute to the spread of extremist ideologies are required. Since this topic is thoroughly emergent, it is necessary to explore implications for the broader role of AI in online radicalisation and thereby identify new insights or perspectives to address the problem. Secondary data collection forms the base of the research design and uses existing sources of information to develop a complete understanding of information. These sources consist of but are not limited to case studies, academic literature, government & cybersecurity reports, and open-source intelligence (OSINT) among open-source platforms. The research is able to focus on secondary data which utilise existing findings of the evolution of AI in radicalisation efforts that help set up a strong base for analysis. The strength of this approach lies also in the fact that it enables one to look at many pieces of data and evidence on impact, which provides a better understanding of AI’s contribution to extremism. Finally, because this research has been implemented with an exploratory approach, it permits the flexibility to examine many different aspects of the issue, including the psychological side of radicalisation, the technological mechanisms of AI-driven manipulation, and geopolitical and security implications. Secondly, this methodology helps in identifying gaps in the existing literature and figuring out the areas where future research would be conducted. Data for this research is collected from multiple sources and includes open source intelligence (Open source intelligence, OSINT) which refers to publicly available materials open to anyone to analyse, social media platforms, government publications, and cybersecurity reports. The researcher has access to information that is freely shared by individuals, organisations, and institutions through the use of OSINT and can use this to gain insight as to how extremist groups use AI-driven tools to aid propaganda and recruitment. As such, social media platforms befitting Twitter, Facebook, YouTube, and Telegram, among others, are particularly relevant given their centrality to the dissemination of extremist content. Through monitoring these platforms, the research is made possible to obtain real time data on how AI is used within radicalisation efforts. Secondary data regarding AI’s role in online manipulation and radicalisation can also be acquired from academic journals and third-party cybersecurity reports in addition to OSINT. What sets these sources apart is a more analytical, that is, empirical research and expert analysis of how AI technologies may be used by extremist groups and how they might be leveraged to influence public discourse. There are studies in the academic literature of AI algorithms, the psychology of radicalisation, and socio-media dynamics, and the cybersecurity reports provide technical information on how AI has been used in AI-driven attacks, which are bot networks, automated content creation, and using machine learning for data mining and manipulation. A number of case studies built on well-documented instances of how extremist groups recruit and spread their ideology (such as ISIS’s use of artificial intelligence to recruit and radicalise individuals) are also included in the data collection process. Like all of the other case studies, these provide real-world examples of the application of AI to extremist activities and concrete examples of how AI technologies can be used for the nefarious impact discussed throughout this paper. Moreover, social media data will be observed on how AI-driven content circulates in online communities. Included in this data is the analysis of algorithmic recommendations on extremism, created or automated posts about extremism and the spread of extremist narratives. The dissemination and its impact on user engagement and interaction will be tracked with the help of social media monitoring tools.

Decoding digital extremism, advanced analytical frameworks for AI-era terrorism research

The data would be analysed through content analysis and network analysis so as to analyse the spread of radical content across digital networks. Content analysis is a systematic approach to analysing the text, images, and videos created by extremist groups. It thus makes it possible to identify common themes, strategies, and rhetorical techniques employed by the AI in manipulating the reader to witness this disinformation campaign (Mughal et al., 2023). It refers to coding and categorising different kinds of content (like videos, blog posts, social media updates and images) to reveal the types of propagation of extremist ideologies. The language of the content, the appeals to emotion made in the content and the particular tactics of engagement employed will be analysed. Utilising network analysis, the map of the spread of extremist content by the digital networks will be made together with content analysis. Network analysis allows one to study how information flows through the different platforms and across online communities of radical content and how the origins of radical information spread to larger audiences. The benefit of this method is that it gives a visual view of the flow of information, along with identifying key influentials, information hubs and critical nodes of the network. Through network analysis, AI algorithms are shown to help the viral spread of extremist content through the analysis of aspects such as social media interactions, user behaviours, and supply of amplification messages for radical messages (Hossain, 2015). Furthermore, the analysis would be based on psychological implications and the use of behavioural science theories for the understanding of AI driven content manipulation tactics. With the help of behavioural science, one can uncover how people’s attitudes and behaviour are led by external stimuli such as emotive content and convincing messages. This paper will explore how AI technologies employ cognitive biases like confirmation bias and social proof to make the usage of radicalising content more effective (Shaban, 2020). This research aims to gain insight into the psychological mechanism by which AI creates radicalisation through understanding how such AI technologies are able to attract vulnerable persons to their ideological fold.

Analysis and discussion: AI algorithms in radicalisation

The extremist content is now driven by AI-driven algorithms, which act as the main contribution to the radicalisation of the person by amplifying the ideologies over digital platforms (Parvez and Hastings, 2022). Whilst these algorithms were aimed at increasing user engagement, it has the unintended consequence of ranking content that elicited responses, usually angry, fearful, or hateful emotions, including this extreme and/or violent content (Hollewell and Longpré, 2022). When this is happening in the regulated platforms and now playing major roles in today’s social media platforms like Facebook, Twitter, and YouTube, and even in encrypted platforms such as Telegram, where extremist groups can leverage the use of AI to target individuals based on their emotional vulnerabilities and personal data; this issue becomes even more concerning. In Kashmir, Afghanistan and Syria, terrorist groups have started to take recourse to AI and social media algorithms to spread their beliefs of recruitment, radicalisation and propaganda in this region. For instance, in Kashmir, insurgent groups have employed Facebook and WhatsApp, amongst others, to convey their message of political violence to local youth through personalised content (Clubb et al., 2021b; Olomojobi and Omotola, 2021). The Taliban has used platforms such as Telegram to deliver its messages in Afghanistan, making use of AI algorithms to enhance the delivery of content as per engagements. These platforms are being exploited by these groups to push their radical content against individuals who ever exhibited an interest in subjects of violence or political unrest. Another example of such amplification can be seen in the case of the Syria conflict as well (Jafa, 2005). ISIL groups, among others, recruited fighters from all over the world by using social media platforms such as Twitter and Facebook to tailor-made content to vulnerable individuals by using demographic data and psychological profiles. AI technologies in these contexts make recommendations for ever more radical videos and materials, leading to a loop of further radical video and materials consumption (Clubb et al., 2021a). The case studies of these situations highlight ways in which AI helps in propagating radicalisation through the technique of amplifying extremist messages and targeting vulnerable individuals across regions with insurgencies in progress.

Digital seduction: how artificial intelligence targets vulnerable recruits

With the use of personal AI-driven campaigns, terrorist organisations are able to employ the same recruitment tactics against vulnerable individuals by using AI to draw those individuals in with a large range of demographic, psychographic and behavioural data (Fernandez et al., 2019). Here, if any, is when this capability shines through, especially in the conflict zones of Kashmir, Afghanistan, and Syria, where terrorist groups take up on the local grievances and bring recruits into their fold. The effectiveness of AI-driven recruitment campaigns in Kashmir is more pronounced since the region is highly affected by political unrest and a history of conflict, leading to feelings of being disenfranchised or alienated from mainstream society. The insurgent group in Kashmir, Hizbul Mujahideen, has been exploiting Facebook and WhatsApp to target young Kashmiris (Oktavianus and Davidson, 2023). AI tools can analyse online behaviour and interactions of people to find out those individuals who might be susceptible to extremist ideology. Once identified, they are identified and bombarded with personalised messaging pegging on to their frustrations pertaining to the political and social situation in Kashmir (Monaghan and Molnar, 2016). Through the use of AI, these tailored campaigns are very effective at radicalising people who feel alienated from the state and society. In Afghanistan, the Taliban has used Telegram to recruit fighters and recruit supporters worldwide, and AI is also being used to identify potential recruits there. The Taliban use AI algorithms to segment their audience by finding people who might be inclined to their cause on the basis of online behaviour and search behaviours (Asongu et al., 2019). Personalisation of the messaging by the Taliban allows them to pepper the narrative into peoples’ emotional vulnerabilities of injustice, oppression, and resistance, counter the hegemony and burnout. This allows the recruitment process to be more efficient because the content is delivered specifically to the psychological state of the target audience, which increases the probability of engagement. Extremist groups like ISIS in Syria likewise embraced AI to escalate their recruitment drives, focusing on people who had issues with the socio-political ambience. Because ISIS used AI-enabled bots on Twitter, Telegram and other social media, they could select to interact with individuals, creating a sense of community, belonging and empowerment for people who might not otherwise be included within this community (Thomas-Evans, 2022). In addition to spreading ISIS’s violent ideologies, these AI-backed systems facilitated recruitment through automated responses and personalised content with the users.

Manufacturing consent through AI: propaganda dynamics in digital extremism

The greatest abomination is AI being used to propagate and distribute personalised propaganda in radicalisation. Through the processing of large datasets, AI systems allow extremist groups to create content that appeals to the emotional, more so than on the ideological level, with the targets, which means that they become more effective and reach a greater audience with their propaganda campaigns (Kumar and Taylor, 2024). Moreover, this capability is particularly visible in areas such as Syria, Afghanistan, Kashmir, etc., where social media and encrypted messaging platforms become the main channels for disseminating terrorist messages (Lombardo, 2020). The ISIS organisation in Syria particularly complimented the use of AI to develop propaganda for their organisations to present their violent activities as justifiable and righteous. AI content generation was utilised by ISIS to produce hundreds of thousands of videos, still images and articles which were distributed to platforms such as Telegram and Twitter (Proroković and Parezanović, 2023). These platforms leveraged the use of AI algorithms to disseminate the quick ISIS content across the global networks and spread the message to very large audience. Through use of AI, ISIS shaped their messages so that it presented itself as a successful, organised group that promised meaning and importance to the victims. The Taliban has been able to even produce personalised propaganda in Afghanistan through the use of AI tools. This is the result of the combination of the data on emotional states, political views, and social behaviour of individuals by the Taliban’s narratives of resistance of foreign occupation (Treleaven et al., 2023). This allows the Taliban to continue projecting their ideological link and recruit new members by sending these tailor made messages because these tailor made messages send deeper chord of local grievances. This is due to the fact that the group can create AI content on Telegram platforms, enabling it to sometimes create a semblance of a constant presence in spreading its radical messages globally and also mobilising the sympathisers locally and internationally (Aydogdu, 2013). So have extremist organisations in Kashmir exploited AI-driven social media algorithms to spread their propaganda. Often, these groups recruit from the disillusioned youth of the region, attracting them with ideologies to which their perception of injustice and political oppression leads them (Saxena, 2024). Groups in Kashmir can exploit the targeting capabilities of social media and spread content intended to provoke an emotional response among someone, either anger at state repression or support of armed resistance. These groups were able to recruit new supporters through the disguised and AI-driven spread of such messages across the virtual space via tailoring the content creation (Rana et al., 2022).

Legal frameworks and ethical dilemmas in AI-driven counterextremism

AI used for radicalisation and extremism presents some ethical and legal questions, such as emotion manipulation, deceptive content generation and intrusion of private life. Emotional manipulation of vulnerable people is one of the main ethical issues (de Londras, 2018). AI-driven systems can be used to tap into all that is psychologically weak in a person through the use of emotive content to manipulate the ways they act and think (Cybenko and Cybenko, 2018). This will inevitably create issues regarding the morality of employing AI in this manipulative manner when it results in harm, including violence and radicalisation. In addition, there is another important ethical concern about deepfakes and other AI-generated deceptive content. Exploiting the deepfake technology for the purpose of manipulating the video and the audio in order to show a false representation of events or an individual has also been damaging for extremist groups in influencing public perception (Matthes et al., 2023). It has also been used very recently to create deepfakes of political figures in regions like Syria and Afghanistan, supporting violence and extreme ideologies that terror groups are aligned with to further their mission goals (Barnes, 2022). Here, technology is challenging the doctrine of truth in media, where people struggle more and more to tell the difference between real content and manipulated material. Regarding the use of radicalisation, the law must come to terms with AI for these challenges. The current laws were not made to deal with the very distinct risks of AI-powered content, and the development speed of AI technologies is also beyond the pace at which lawmakers can regulate them. In addition, AI-enabled radicalisation is increasingly crossing borders in nature, and thus, traditional national frameworks are frequently powerless to fight it (Sánchez and Ruiz, 2020). To give you an idea, radical content stemming from AI in Afghanistan or Syria can very quickly be distributed around the other parts of the globe, avoiding existing rules and regulations that apply locally. All this emphasises the necessity to involve international cooperation in the creation of new legal standards for coping with complicated AI aspects in the context of radicalisation and terrorism (Feldstein, 2023).

Policy imperatives, technological solutions in extremism prevention and recommendations

There should then be a set of countermeasures and recommendations developed around the risks of radicalisation through the means of AI. Firstly, we require, in the first place, the need of AI content moderation systems based on AI that will automatically detect and philtre out the extremist content (Pelzer, 2018). That is why the AI systems built for these purposes also need the capacity to detect implicit hate speech and violent rhetoric as well as more implicit forms of radicalising content including such as code, extremist memes, etc. Adopt these AI driven systems on the opinion of counterterrorism, counter hate speech (human rights) and digital ethics, to balance the prevention of radicalisation with protection of freedom of expression (Anderson and Horvath, 2017). Early detection systems also monitor people’s online behaviour looking for signs that a particular person might be at risk of radicalisation so that the authorities can intercept that person before they reach the end of the trip of radicalisation route. Examining the patterns in online conversations, content and interactions, such as made it possible to identify at risk people, to trigger an automated action sending a prompt to a digital literacy campaign, or to reach out to disengage at risk individuals from extremism (Korner and Bottcher, 2019). Even in the global sphere, international cooperation is equally important for how to approach AI related to radicalisation. Governments, tech firms and international organisations will all need to collaborate to define these (regulatory framework around AI usage) and therefore there is a need to place them. In addition, judging the moderation of global standards, watching the use of AI in the activities of extremism, and guaranteeing that AI technologies are ethically developed and deployed in this form (Goldstein et al., 2024). For this to happen will require all parties to collaborate to come up with cross border agreements regarding the sharing of information and resources to address the misuse of AI. Finally, although there are many advantages of AI in a variety of businesses, the use of AI in radicalisation and extremism is a growing threat that we urgently need to address (Downing, 2023). With the development of refined countermeasures, regulations, and cross-country interaction, dangers from AI in radicalisation and terrorism can be brought down, securing both security and human rights in the advanced age.

Framework for AI regulation in counterterrorism: international collaboration

However, as extreme organisations adopt Artificial Intelligence (AI) in order to radicalise and recruit and to posture and disseminate violence, the response to this challenge must be multinational (Borelli, 2023). Counterterrorism efforts must not only depend on national security agencies but the work must be done in conjunction with the governments, technology companies and international organisations across the globe (Nirupama, 2023). AI-enabled forms of radicalisation across countries; the digital world is very complex and needs to be regulated and monitored by a united body. Shared frameworks to combat AI-driven extremism have to be established globally, and international cooperation is needed for that purpose (Conway and Macdonald, 2021). Tech companies must collaborate with governments, especially those in areas prone to extremist threats, to more strictly moderate content and also be more vulnerable in regards to their algorithms (CB Insights, 2018). Facebook, YouTube, and Twitter are directly involved in preventing its propagation within their platforms. Governments need to work with them to develop and enforce policies that reduce the chance of AI being used to proliferate radicalising ideas without presenting the risk of stifling freedom of speech and privacy (Rubin, 2022). International agencies such as the United Nations and Interpol must serve as mediators of global cooperation by establishing communication between the latter and facilitating dialogue among them, associated coordination of information exchange, and the development of global rules on AI use in counterterrorism (Piazza, 2022). Therefore, Interpol could assist the member states in creating global databases of flagged extremist content, shared threat intelligence of AI-algorithmic-driven radicalisation methods and a coordinated response to cyber terror attacks (Winter et al., 2020). Moreover, the United Nations could promulgate international norms and principles of good practice of how to counter the risk of utilisation of AI for violent extremism and assist member states to effectively conform with these standards and take into account domestic sovereignty (Dean et al., 2012). Also, the collaboration must extend to the sharing of best practices in the development of AI-based counterterrorism technology, which will allow the nations to learn from each other as they combat evolving threats of AI in radicalisation. Governments and organisations can use their resources and expertise to access a collective effort in tracking, monitoring, and combating AI-driven extremism on a worldwide scale (Magu et al., 2017). The speed of technological advancement for AI technologies raises the question of responsible implementation of AI in counterterrorism. It becomes important that the AI frameworks like AI ethics should be developed robustly so that the use of AI in the fight against radicalisation and extremism can be done responsibly. This addressing must comprise not only the technological and security aspects of AI but also how AI is to be brought about in a decent society that respects privacy, allows for freedom of speech, and does not leave a tired sexist algorithm alone to decide about its freedom to choose (Shahid et al., 2024). In order for AI use in counterterrorism to be ethical, AI technologies—especially those on sites where social media are deployed—need to be designed and used in compliance with human rights and in a way that does not violate civil liberties. In particular, AI systems that are used for content moderation (like Twitter) or behavioural profiling (like Facebook) must not violate individual privacy rights and not have excessive impact on some particular groups by using biased data (McKendrick, 2019). For example, there should be transparency in the workings of AI tools for surveillance or online monitoring, for instance, so that data is collected and analysed ethically, following the laws and that clear guidelines on how the collected personal information is handled. In addition, ethics frameworks for the use of AI technologies ought to focus on the accountability of tech firms and governments in the usage of AI technologies (Fraiwan, 2022). To guarantee that AI systems do not, by accident, maintain biases or radicalise people as a result of bad design, these frameworks must entail it should perhaps require transparency when algorithms are designed and the data that they are using. Policies that the digital platforms should adhere to must include clear standards for the ethical use of AI, otherwise, they will be held liable for misuse of their AI technology in these platforms. Applied to practice, this would translate to the alignment of ethical guidelines with platform policies that cover content moderation such that AI would be used conscientiously to detect and remove extremist content without infringing on legitimate discourse (Rajtmajer and Susser, 2020). Therefore, the establishment of AI ethics frameworks aimed at such fields as counterterrorism should be a dynamic process that continues to develop and adapt in light of the growth of computer science technologies and changes in the threat situation. The international human rights standards should be periodically reviewed and adapted to ensure that AI frameworks are in line with global international human rights standards and the most recent AI technology (Sudmann, 2019).

Artificial intelligence in extremism prevention and detection

In order to address growing AI-driven radicalisation, AI driven counter extremist tools are needed that can detect radicalisation in its early stages as well as identify channels of recruitment and the spread of extremist propaganda. Such tools will be developed based on the use of machine learning and other AI techniques to look for radicalising behaviour and extremist content on digital platforms (Kumar and Taylor, 2024). Preventing radicalisation from progressing requires its early detection because, at this stage, intervention can be performed when people are not so steeped in extremist ideologies. The language, content, as well as shifts in sentiment on the platforms, can be used to train machine learning models to detect patterns of radicalising behaviour, even when radicalisation (KhosraviNik and Amer, 2022). Thanks to analysing big amounts of data, AI systems can raise red flags to suspicious instances of radicalisation in real-time, giving law enforcement agencies and counter-extremism organisations a chance to intervene as soon as possible. More specifically, such tools may be particularly effective in monitoring youth engagement due to youth’s greater susceptibility to extremist messaging and because of the heavy amount of digital media that youth are being exposed to. Apart from radicalisation detection, AI could be useful in identifying the recruitment channels extremist groups are using in their recruitment efforts (Metaxas, 2020).

For instance, antioxidant tools that operate on the principle of AI power go through communication patterns on encrypted messaging tools, such as Telegram and WhatsApp, where lots of extremist groupings join up to recruit. Natural language processing (NLP) and sentiment analysis, AI systems can determine the key phrases, keywords or themes used when recruiting and tell the company what to look out for and when to carry out further investigations because of what’s being said. They can be incorporated into intelligence-sharing platforms such that relevant information is passed to the correct authorities for actual action (Revell, 2017). A second important function of AI assisted counter extremism tools is helping to monitor the spread of extremist propaganda on digital networks. AI algorithms can render a graph of how extremist content propagates through social media or other online spaces and who are the influencers and key nodes among them (Aistrope, 2016). In addition to this, security agencies can use this type of network analysis to analyse the structure of extremist groups and their online presence and be able to identify key individuals who spread radical ideologies (Tian et al., 2022). These same tools can assist in detecting deepfakes and fake news (among other things), as terrorist organisations now apply them to manipulate public opinion, recruit followers, etc. Furthermore, some of these early warning systems can improve the abilities of current counterterrorism measures (Zenn, 2023). Such systems employ predictive analytics and machine learning algorithms to forecast threats that could be escalated in future by extremist actions by spotting such patterns in the behaviour that presages such actions. Through ongoing data analysis from several sources such as social media platforms, news outlets and online forums, these systems may alert authorities on possible acts of terrorism, giving them enough time to prepare and respond adequately. All in all, the development of AI-driven counter-extremism tools shows great promise in identifying, monitoring, and minimising the dangers of AI in global radicalisation and extremism (Pandey et al., 2022). The use of advanced machine learning models and network analysis methods gives security agencies and digital platforms a constructive role in monitoring and disrupting extremist activities. To succeed, though, these tools will have to be deployed in conjunction with robust and effective international cooperation, ethical frameworks, and regulation, either by private or, more properly, national agencies, to make certain they are applied for and to countering extremism only (Bazarkina, 2021).

Case studies: illuminating AI’s role in extremist radicalisation and counterterrorism

Case studies: ISIS’S use of AI-driven social media algorithms for recruitment

The Islamic State of Iraq and Syria (ISIS) have led the pack when implementing artificial intelligence (AI) technologies to enhance its recruitment and propaganda (Ceron et al., 2019). Their plan is primarily to exploit the recommendation algorithms computed by AI on the social media platforms that people around the world use universally, including Telegram, YouTube, and Facebook. Those sites employ a potent machine learning algorithm that is guided to achieve the highest number of user interactions by customising user content feeds to their tastes and based on what they have encountered before (Desa et al., 2022). The algorithmic logic that ISIS is using is that of producing emotionally charged, provocative and personalised content that is likely to attract individuals who feel disenfranchised or discontented with their socio-political situation. By profiling behaviour with the help of AI, ISIS targets users who are vulnerable to extremist content with individualised videos, messages, and interactive content. As one case in point, militants have created AI-optimised message chains that advance radical ideologies, using information about the user to adjust intensity and themes dynamically. It is an adaptive messaging method that uses AI with the ability to process large volumes of data in a short time, which has brought the rate of conversion to dramatic levels (McElreath et al., 2018). People have turned passive audiences into active recruits. Scholarly studies have shown that the radicalisation process is mainly fueled by the amplification of radical content through platform algorithms, which is a crucial factor contributing to the process and making it more severe. The closer the content is to individuals and the more interesting it becomes, the more likely they will pass through different levels of radicalisation as envisaged in the Staircase to Terrorism model by Moghaddam. European extremist groups are therefore recruited in large numbers by an algorithmic content amplifier (Awan, 2017). This case highlights significant difficulties in counterterrorism. Current systems of content moderation, which frequently rely on human effort or rules, struggle to keep up with AI-generated or AI-optimised extremist content that is constantly evolving to avoid detection. Besides, AI transparency advocates believe that the AI operations should be made more transparent in social media, but these proposals have been opposed because of property considerations (Tulga, 2022). Finally, the strategic manipulation of AI algorithms by ISIS underscores the necessity of multi-stakeholder cooperation between tech companies, governments, and academia so as to establish robust, ethical and scalable AI governance frameworks that will efficiently counter the two aspects of AI technologies of dual-use (Cunliffe and Curini, 2018).

Al-Qaeda’s adoption of generative AI for propaganda content production

The radical propaganda machine of al-Qaeda has gradually begun to incorporate generative technologies of artificial intelligence to create synthetic media, such as deepfake videos and AI-generated audio speeches (Rudner, 2017). The use of generative AI systems, including text-to-speech and image generation networks, allows the group to produce some of the most believable yet fake media that allegedly features the comments or recommendations of high-ranking personnel. The benefits to the operations are two-fold: to reduce the limitation of resources, first, generative AI automates propaganda creation that would otherwise be performed through human effort, therefore, producing more and more content of a higher volume and more often; second, the synthesised content is difficult to debunk, and it confuses both enemies and target audiences (Choi et al., 2018). Intended to inspire aggression in the group, deepfakes, by example, can be used to recreate speeches by leaders, giving the impression of a sense of cohesion and urgency in the group. This phenomenon has been identified by the academic community as the hallmark of the new menace of synthetic media in the face of terrorism (Zvosdetska et al., 2023). Hyper-realistic AI content threatens traditional counter-propaganda strategies that are based on source validation or eyewitness validation. This has led to studies on AI-based methods of detection systems to detect the digital artefacts left by generative algorithms (Ingram, 2016). Also, there is a geopolitical implication of Al-Qaeda’s engagement in synthetic propaganda. The group utilises AI-generated content based on regional dialects and cultural icons, targeting specific segments of the population in various nations. The spread of such content is further exacerbated by online transnational radical networks, which exacerbate global security issues (Weimann, 2012). The suggested countermeasures include international coordination of intelligence exchange on the use of synthetic AI, platform-based AI screening mechanisms that are sensitive to deepfakes, and rights-compliant legislation that requires AI-generated information to be disclosed (Perešin, 2014). This case study highlights the pressing need to come up with adaptive technological countermeasures and legal frameworks in tandem with the continued weaponisation of AI by extremists.

Taliban’s AI-enhanced messaging on encrypted platforms

Since the Taliban reclaimed power in Afghanistan, it has successfully used AI-powered applications embedded in encrypted messaging applications, particularly Telegram, to optimise recruitment and propaganda efforts (Sahel, 2020). Their approach is based on the privacy provided by encryption and the scalable personalisation provided by AI. As chatbots are AI-driven, they replicate one-on-one conversations, and the targeted recruits can be engaged in an automated, yet context-aware, conversation (Tabish et al., 2022). These chatbots rely on natural language processing (NLP) models to process user responses and generate answers in the form of human-like responses, which include ideological messages, logistical information, and personal stories. The AI-enhanced method employed by the Taliban represents a significant step forward compared to traditional recruitment methods, which rely solely on face-to-face interaction or manual recruitment methods (Bernatis, 2014). With automation, the group reduces the risk to personnel and increases its reach and coverage, ensuring continuous engagement even in disputed regions. This can be defined as a technological innovation because it allows tailoring the content based on individual user profiles and a set of behavioural patterns monitored in near-real time, and the algorithm for popularising Taliban content is based on engagement-centred ranking of content (Drissel, 2015). Network analysis reveals that there are nodes that are concentrated, serving as the most influential propagators of the Taliban story, and that AI further enhances exposure. The AI-enabled messaging mix protocol of the Taliban combines localised socio-cultural expertise with an AI incorporation, and is a hybrid type of recruitment scheme that is exceptionally resistant to other counterterrorism strategies (Mozur and ur-Rehman, 2021). This requires sophisticated methods of detection that involve the use of algorithmic detection and subtle cultural and language study to predict radicalisation waves (Bahar, 2020). Ethical issues also arise in this case, as encryption also implies privacy; at the same time, extremist AI actions are not adequately monitored. Creating a balance between these opposing needs is an inherent challenge for policymakers and those involved in counterterrorism.

Far-right extremist movements’ use of AI for memetic warfare

Far-right extremist movements in Western democracies were the first to utilise AI-generated memes and images as a key element of their ideological warfare on the internet. With the help of artificial intelligence (AI), such as generative adversarial networks (GANs), these groups produce memes in large volumes, designed to be emotionally and politically relevant to specific demographics (Holt et al., 2022). AI shortens the time spent on the content production process and enables the rapid promotion of exceptionally personalised, symbolically loaded, and linguistically modified memes in social media, such as Twitter and Gab. The memes frequently use coded language, dog whistles, and visual symbols that strengthen the sense of in-group and enhance the aggression toward the perceived enemies (Karpova et al., 2022). The memetic warfare approach supports the development of echo chambers with the help of AI-driven engagement optimisation algorithms. Repetition of the same patterns and stories, with the help of AI-based curation, makes people more polarised and accept extremist wording as a normal phenomenon in the digital world. Social network analysis reveals dense groups of users who intensify such content, and the central nodes may be both creators of the memes and critical agents of their spread (Klinger et al., 2023). This content is viral, which helps achieve recruitment success by reducing ideological barriers and reaching politically indifferent youth. The case demonstrates that counter-narrative activity, based on the use of AI tools to respond quickly, counteract, and offer digital literacy training, is required. It also focuses on the ethical requirement to strike a balance between what is shared on the platform and the freedom of speech, which must be approached carefully with community standards in mind in the context of AI-managed meme ecosystems (Am and Weimann, 2020).

International counterterrorism efforts utilising AI for early detection

The US, the UK, and several European countries have deployed AI-based technology to track online radicalisation and recruitment efforts at the initial stage. These kinds of systems are based on machine learning classifiers, which are trained using large volumes of extremist information and indicators of their behaviour to attract the attention of possible human analysts to suspicious activity (Al-Rikabi, 2021). One of the technologies would be software that analyses linguistic data to find radical vernacular, which is able to track network actions in order to determine the rising number of extremists. Additionally, there would be multimedia content scanning software that determines AI-generated or manipulated propaganda (Monar, 2015). As a case in point, the UK Home Office tested the use of AI-powered social media scanning to identify potential radicalisation and integrated it with local law enforcement intervention strategies. Although these technologies are promising, they raise critical ethical and legal questions regarding privacy, data protection, and bias in the algorithms (Saadat, 2020). To avoid the imbalance of targeting or incorrect labelling of minority groups, the surveillance programs should be in accordance with the national and international human rights standards. The effectiveness of these detection tools is constantly measured by feedback loops, which improve the algorithms to minimise false positives and become more culturally relevant. This case highlights the need for interdisciplinary cooperation among technologists, legal scholars, and social scientists to develop responsible AI counterterrorism models (Burchardt and Gulati, 2018).

Extremist use of social media for terrorist propaganda in Kashmir

Social media has become a weapon that separatist and terrorist groups use in the crowded socio-political stage of Jammu and Kashmir by spreading propaganda and getting supporters via WhatsApp, Facebook, and Telegram (Asatryan and Kalpakian, 2023). The region has highly remained disturbed and therefore, traditional communication is challenging to achieve; digital platforms offer a powerful alternative for messaging and mobilisation. Such groups utilise AI-driven amplification strategies, using bots and optimising content based on algorithms, to achieve maximum reach and engagement (Gabel et al., 2022). Videos that promote militancy, appeals that are crafted in the context of political victimhood, and calls to sell arms are carefully disseminated to the local people and the Kashmiris in the diaspora groups. Analysis of social media posts using sentiment analysis reveals that the content is highly emotional, designed to inspire feelings of collective grievance, perceived injustice, and identity affirmation (Sufi, 2023). Messaging is also refined using AI tools to address various linguistic and cultural groups in Kashmir, making it more resonant and compelling in recruiting. Counterterrorism officials are torn between fighting such AI-powered digital campaigns and civil liberties, without creating divisions among local people (Speckhard and Ellenberg, 2021). They involve the introduction of AI-driven surveillance devices, along with community outreach and digital literacy interventions (Olomojobi and Omotola, 2021). The case also emphasises the utmost significance of the relationships between developed AI technologies and local conflict processes, and describes the necessity of implementing both situation-specific and ethically grounded counter-radicalisation policies in Kashmir and other conflict regions (Asongu et al., 2019).

Limitations of the study

However, this research nevertheless has a number of limitations despite its comprehensive approach. The problem is also about the lack of access to classified information or real-time extremist activity data. However, because some extremist groups and their activities involved in radicalisation campaigns and recruitment efforts are sensitive, a large amount of the real-time data is classified or has been restricted to security agencies and intelligence organisations. Therefore, the research is very dependent on data made publicly available since it does not claim to have a measure of all the ways AI has been utilised in radicalisation or the most recent tactics that extremists use. Moreover, limitations of potential biases in data sources, especially from government agencies and organisations with some particular agendas, maybe another limitation. For instance, cybersecurity reports are typically produced by agencies which are charged with the task of monitoring and controlling extremist activities. However, these reports offer useful information in regard to the technical aspects of AI-driven radicalisation but may reflect the priorities or concerns of the particular organisation producing the report. In the same way, biases might be found in academic literature based on where the authors are affiliated institutionally or in the political environments in which research is conducted. The researcher must be careful when interpreting the findings from such sources as they are biased. Moreover, AI technologies and extremist tactics are racing to catch up with one another, so the research’s findings may not represent the most current developments in the ever-changing landscape. As AI grows, there will be more manipulation and radicalisation online, and more effective countermeasures than what we can foresee will be needed to combat them, and only through continuous monitoring and adaptation of them will we be able to ensure some level of success. The reality that this is so rapidly changing presents issues for this research, as it cannot be fully prepared to foresee future trends and developments in the use of AI for extremist uses. The final limitation is based on the fact that the research methodology relies on secondary data and may not provide a complete picture of how AI is currently being used in radicalisation, as much of the information is not publicly accessible or underreported. Because of this, the research is limited to available data and can, therefore, not show the full effect that AI has on extremist recruitment and propaganda. Overall, this is a qualitative research, an exploratory research method, using content analysis, network analysis, and psychological analysis to examine the role the Act played in radicalisation, recruitment, and propaganda. Based on secondary data collection, the paper discusses how AI is being used in extremist activities, including case studies, social media data and cybersecurity reports. While these tools are limited, the research provides some insights into how extremist organisations are using AI and serves as a foundation for future work in this critical issue and as a first line of defence against the deployment of these tools by extremists.

Conclusion: addressing the dual-edged sword of AI in counterterrorism

One of the more urgent challenges in the digital age is the role that Artificial Intelligence (AI) is taking on in processes of radicalisation, recruitment and propaganda. In addition to making extremist activities more efficient and massive, AI also opens up new ways for terrorist organisations to lure the vulnerable and transmit their extreme ideas. Extremist groups have been able to extend their reach, influence and manipulation by using tools such as social media bots, personal content algorithms and other automated content generation. Through these, they have managed to do so with less resistance from the current counterterrorism efforts in place. However, AI is also a powerful tool; its capacity to tailor messages, predict behaviour, and amplify narratives that might be strongly influenced by extremists is something that has never been done before. One key result of this analysis is that AI algorithms, and especially those used by social media platforms, have inadvertently helped radicalisation to be amplified because they marketed emotionally loaded and divisive content. They prioritise things that people react to, and therefore, the more emotion there is involved in the media, the more they see there (a place for outrage or fear). Terrorist groups have used complex AI technologies in war zones such as Kashmir, Afghanistan and Syria to spread propaganda, recruit new members and even inspire violence in other parts of the world. For instance, extremist organisations in these regions exploit Facebook, YouTube, and Telegram to reach young people with grievances related to political oppression, economic suffering or social alienation. Through the nature of their work, these data-driven AI systems give these groups the ability to specifically create very targeted messages very explicitly and very personal and an almost direct jab at what a vulnerable population is experiencing, causing them to feel alienated and exasperating their receptivity to extremist ideologies. It also points out how AI assists in recruitment, making it more efficient and more difficult to detect by using data provided by the profile, manipulation of the psyche, and personalised messaging. Extremist groups use AI tools in order to figure out individuals who are more likely to be radicalised and then produce content suited to the emotional and ideological needs of those individuals. AI-driven recruitment efforts hold their power in marginalised communities where people are already turned away or outside the privileges of society. AI is being used by extremist groups to take advantage of these vulnerabilities by providing a belonging and a purpose to those they sense as disconnected from society. Likewise, propaganda has adopted AI methods for its spread, and extremists have taken advantage of advanced technology to create content that is more exciting, believable, and extensive than what we were accustomed to with the more normal methods. Regarding AI, as much of content generation, especially in the form of deepfakes, bots, and automated videos, terrorist organisations were able to magnify their voice across digital space on a scale never seen before. Although these AI tools are driven by extremism, it becomes very challenging to trace, moderate and limit the circulation of dangerous content by authorities simply because it is driven by extremist ideology and reaches a big chunk of the global audience. Radicalisation and extremism are a significant challenge in the field of ethics and the law relating to AI. Any capability by AI technology to take advantage of an individual’s psychological vulnerabilities is demonstrably unethical and should be an area of concern for manipulation, privacy, or consent. The use of deepfake technology and the creation of deceptive media has added serious complexity to this already complicated problem because now people can be so easily deceived and misdirected. Trust can be destroyed against real media sources. Current frameworks are not legally equipped to deal with the matters raised by AI-driven extremism. Many laws were developed before there were such sophisticated technologies and therefore, cannot control and regulate the global, cross-border nature of the abuse of AI in terrorism. However, due to the absence of a unified international regulatory approach, extremist groups can obviously use the gap left by national laws to breach, making enforcement more cumbersome. Given this, it is essential to set up wide-ranging AI regulation policies. One of these policies will be critical to ensuring a proper balance of security and privacy. AI has a huge potential in combating extremism in general but, more importantly, in the detection of radicalisation behaviours and monitoring of radical/extremist content, yet strong regulation is required to restrict the violation of privacy rights and preclude counterterrorism measures against fundamental freedoms. Moreover, there will also be the need to integrate ethical guidelines into the design and launch of AI systems so as to prevent risks associated with exploitation and misuse. On the one hand, governments, together with technology companies should develop meaningful measures ending up with unbiased and non-extremist work with transparent and accountable AI systems. Another critical part of the answer to AI-fuelled radicalisation is the creation of AI-fuelled counter-extremism tools. Knowing early warning signs of radicals is important in spotting them in time, and machine learning models can monitor changes in language, sentiments and online behaviour to detect early signs. They should be able to find recurrent patterns indicating an individual’s exposure to extremist content. They may alert authorities or intervention programmes as early as an individual is at risk of being fully radicalised. Moreover, AI-based early warning systems are capable of giving real time data to the levelling enforcement so they can respond to the dangers before they rise up. For disruption of extremists’ recruitment channels and dismantling of radicalising networks, it will also be crucial to be able to track extremist propaganda across digital networks using AI. Finally, fighting back against extremism aided by the forces of AI cannot be done without international support and multi-stakeholder involvement. Governments, tech companies, international organisations as well as civil society need to engage together to develop and enforce policies on how AI should be utilised in the digital space. Therefore, this collaboration should be about developing global standards for the ethical application of AI, not only in counterterrorism but more broadly, and making sure this application in counterterrorism is both rights-respecting and effective. Stakeholders can share intelligence, best practices and technological resources to ensure there is a one through approach in the fight against extremism in the digital age. However, with the progressive advance of AI technology, it is now mandatory for counterterrorism strategies to advance as well. The fight against extremism and radicalisation facilitated through the AI systems is an almost infinite story. Through the development of strong AI-based counter-extremism tools and ethical guidelines from the policymakers, the threats posed by AI to radicalisation and propaganda can be minimised, and a safer digital place can be made for all. Finally, only concerted action by the world’s governments, tech companies and organisations can see us tackling the growth of this new form of extremism. The secret to success will be in us using and regulating AI in the interests of society, not in the hands of those who will do it damage and divide.

References

• 1

  AistropeT. (2016). Social media and counterterrorism strategy. Aust. J. Int. Aff.70, 121–138. doi: 10.1080/10357718.2015.1113230

  • CrossRef
  • Google Scholar

• 2

  Al-RikabiM. K. (2021). The role of the Interpol in the counter-terrorism in Iraq 2003-2017. E-Bangi J. Soc. Sci. Humanit.18, 187–201.

  • Google Scholar

• 3

  AmA. B.WeimannG. (2020). Fabricated martyrs: the warrior-saint icons of far-right terrorism. Perspect. Terror.14.

  • Google Scholar

• 4

  AndersonB.HorvathB. (2017). The rise of the weaponized AI propaganda machine. Medium.

  • Google Scholar

• 5

  AprinF.ManskeS.ChountaI. A.HoppeH. U.2021. Is this fake or credible? A virtual learning companion supporting the judgment of young learners facing social media content. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics).

  • Google Scholar

• 6

  AsatryanG.KalpakianJ. (2023). The Palgrave handbook of national security. Intell. Natl. Secur.38. doi: 10.1080/02684527.2022.2031664

  • CrossRef
  • Google Scholar

• 7

  AsonguS. A.OrimS. M. I.NtingR. T. (2019). Terrorism and social media: global evidence. J. Glob. Inf. Technol. Manag.22, 208–228. doi: 10.1080/1097198X.2019.1642023,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 8

  AwanI. (2017). Cyber-extremism: Isis and the power of social media. Society54, 138–149. doi: 10.1007/s12115-017-0114-0,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 9

  AydogduU. F. (2013). “Technological dimensions of defence against terrorism” in NATO Science for Peace and Security Series. E, Human and Societal Dynamics. Amsterdam: IOS Press.

  • Google Scholar

• 10

  BaharH. M. (2020). Social media and disinformation in war propaganda: how Afghan government and the Taliban use twitter. Media Asia47, 34–46. doi: 10.1080/01296612.2020.1822634

  • CrossRef
  • Google Scholar

• 11

  BalcerowiczP.KuszewskaA.2022Law and conflict resolution in Kashmir. Amsterdam: IOS Press.

  • Google Scholar

• 12

  BamseyO.MontasariR. (2023). “The role of the internet in radicalisation to violent extremism” in Advanced sciences and technologies for security applications. Amsterdam: IOS Press.

  • Google Scholar

• 13

  BarnesM. R. (2022). Online extremism, AI, and (human) content moderation. Fem. Philos. Q.8. doi: 10.5206/fpq/2022.3/4.14295

  • CrossRef
  • Google Scholar

• 14

  BazarkinaD. 2021. Artificial intelligence enhancing social Engineering in terrorism: the new threats. 3rd European conference on the impact of artificial intelligence and robotics, ECIAIR 2021.

  • Google Scholar

• 15

  BernatisV. (2014). The Taliban and twitter: tactical reporting and strategic messaging. Perspect. Terror.8.

  • Google Scholar

• 16

  BierstekerT. J. (2007). Countering the financing of terrorism. London: Routledge.

  • Google Scholar

• 17

  BlasiakK. M.RisiusM.MatookS. 2021. Conceptualising social bots for countering online extremist messages. ACIS 2021 - Australasian conference on information systems, proceedings.

  • Google Scholar

• 18

  BorelliM. (2023). Social media corporations as actors of counter-terrorism. New Media Soc.25, 2877–2897. doi: 10.1177/14614448211035121,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 19

  BurchardtD.GulatiR. (2018). International counter-terrorism regulation and citizenship-stripping laws-reinforcing legal exceptionalism. J. Confl. Secur. Law23, 203–228. doi: 10.1093/jcsl/kry012

  • CrossRef
  • Google Scholar

• 20

  ButlerG.MontasariR. (2023). “The impact of the internet on terrorism and violent extremism” in Advanced sciences and technologies for security applications. Amsterdam: IOS Press.

  • Google Scholar

• 21

  CalvertJ. B. (2024). Engagement with radical propaganda drives cognitive radicalisation: an analysis of a right-wing online ecosystem. J. Strateg. Secur.17, 24–30. doi: 10.5038/1944-0472.17.1.2160

  • CrossRef
  • Google Scholar

• 22

  CasebeerW. D.ZieglerM.KraftA. E.PoleskiJ.RussellB. (2018). Human performance augmentation in context: using artificial intelligence to deal with variability—an example from narrative influence. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics).

  • Google Scholar

• 23

  CB Insights (2018). Memes that kill: the future of information warfare. Res. Briefs.

  • Google Scholar

• 24

  CeronA.CuriniL.IacusS. M. (2019). ISIS at its apogee: the Arabic discourse on twitter and what we can learn from that about ISIS support and foreign fighters. SAGE Open9. doi: 10.1177/2158244018789229

  • CrossRef
  • Google Scholar

• 25

  ChoiK.LeeC. S.CadiganR. (2018). Spreading propaganda in cyberspace: comparing cyber-resource usage of Al Qaeda and ISIS. Int. J. Cybersecur. Intell. Cybercrime1, 21–39. doi: 10.52306/01010418zdcd5438

  • CrossRef
  • Google Scholar

• 26

  ClubbG. (2019). De-radicalisation, disengagement and the attitudes–behavior debate. Terror. Political Violence. doi: 10.4135/9781473917248.n18

  • CrossRef
  • Google Scholar

• 27

  ClubbG.KoehlerD.ScheweJ.O’ConnorR. (2021a). “Conceptualising de-radicalisation programmes” in Selling De-radicalisation (London: Taylor & Francis Group), 19–40.

  • Google Scholar

• 28

  ClubbG.KoehlerD.ScheweJ.O’ConnorR. (2021b). “Selling de-radicalisation: managing the media framing of countering violent extremism” in Selling De-radicalisation: managing the media framing of countering violent extremism. Amsterdam: IOS Press.

  • Google Scholar

• 29

  ConwayM. (2017). Determining the role of the internet in violent extremism and terrorism: six suggestions for progressing research. Stud. Confl. Terror.40, 77–98. doi: 10.1080/1057610X.2016.1157408

  • CrossRef
  • Google Scholar

• 30

  ConwayM.MacdonaldS. (2021). Introduction to the special issue: extremism and terrorism online—widening the Research Base. Stud. Conflict Terrorism, 1–7. doi: 10.1080/1057610X.2020.1866730,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 31

  CunliffeE.CuriniL. (2018). ISIS and heritage destruction: a sentiment analysis. Antiquity92, 1094–1111. doi: 10.15184/aqy.2018.134

  • CrossRef
  • Google Scholar

• 32

  CybenkoA. K.CybenkoG. (2018). AI and fake news. IEEE Intell. Syst.33, 1–5. doi: 10.1109/MIS.2018.2877280

  • CrossRef
  • Google Scholar

• 33

  Da SilvaC.AmadioN.SargR.DomingoB.BenbouricheM. (2023). A decade of media coverage of the social reintegration of terrorism-related convicts: France as a case study. Terror. Political Violence36, 1064–1082. doi: 10.1080/09546553.2023.2248269,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 34

  de LondrasF. (2018). “Evaluation and effectiveness of counter-terrorism” in Socially responsible innovation in security: Critical reflections. Amsterdam: IOS Press.

  • Google Scholar

• 35

  DeanG.BellP.NewmanJ. (2012). The dark side of social media: Review of online terrorism. Sogeti VINT Research Institute for New Technology, The Netherlands.

  • Google Scholar

• 36

  DesaM. B. M.NasirN. B. C. M.JasniM. A. B.YusofY. B.NordinM. N. (2022). ISIS uses a social influence techniques to induce individuals to become terrorist sympathizers: a review. J. Pharm. Negat. Results13, 5622–5630.

  • Google Scholar

• 37

  DowningJ. (2023). “Social media, security and terrorism in the digital age” in New security challenges, Amsterdam: IOS Press. 109–139.

  • Google Scholar

• 38

  DrisselD. (2015). Reframing the Taliban insurgency in Afghanistan: new communication and mobilization strategies for the twitter generation. Behav. Sci. Terror. Polit. Aggress.7, 97–128. doi: 10.1080/19434472.2014.986496

  • CrossRef
  • Google Scholar

• 39

  FeldsteinS. (2023). The consequences of generative AI for democracy, governance and war. Survival65, 117–142. doi: 10.1080/00396338.2023.2261260

  • CrossRef
  • Google Scholar

• 40

  FernandezM.AsifM.AlaniH. 2018. Understanding the roots of radicalisation on twitter. WebSci 2018—proceedings of the 10th ACM conference on web science.

  • Google Scholar

• 41

  FernandezM.Gonzalez-PardoA.AlaniH. (2019). Radicalisation influence in social media. J. Web Sci.6.

  • Google Scholar

• 42

  FraiwanM. (2022). Identification of markers and artificial intelligence-based classification of radical twitter data. Appl. Comput. Informatics21, 304–316. doi: 10.1108/ACI-12-2021-0326

  • CrossRef
  • Google Scholar

• 43

  GabelS.ReichertL.ReuterC. (2022). Discussing conflict in social media: the use of twitter in the Jammu and Kashmir conflict. Media War Confl.15, 504–529. doi: 10.1177/1750635220970997

  • CrossRef
  • Google Scholar

• 44

  GoldsteinJ. A.ChaoJ.GrossmanS.StamosA.TomzM. (2024). How persuasive is AI-generated propaganda?PNAS Nexus3. doi: 10.1093/pnasnexus/pgae034,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 45

  GrazianiD. C. (2023). International terrorism, radicalisation and technology. Federalismi.It2023.

  • Google Scholar

• 46

  HerathC.WhittakerJ. (2023). Online radicalisation: moving beyond a simple dichotomy. Terror. Polit. Violence35, 1027–1048. doi: 10.1080/09546553.2021.1998008

  • CrossRef
  • Google Scholar

• 47

  HollewellG. F.LongpréN. (2022). Radicalisation in the social media era: understanding the relationship between self-radicalisation and the internet. Int. J. Offender Ther. Comp. Criminol.66, 896–913. doi: 10.1177/0306624X211028771,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 48

  HoltT. J.FreilichJ. D.ChermakS. M. (2022). Examining the online expression of ideology among far-right extremist forum users. Terror. Polit. Violence34, 364–384. doi: 10.1080/09546553.2019.1701446

  • CrossRef
  • Google Scholar

• 49

  HossainM. S. (2015). Social media and terrorism: threats and challenges to the modern era. South Asian Surv.22, 136–155. doi: 10.1177/0971523117753280

  • CrossRef
  • Google Scholar

• 50

  IdahosaS. O. (2017). International terrorism: the influence of social media in perspective. World Wide J. Multidiscip. Res. Dev.3.

  • Google Scholar

• 51

  IngramH. (2016). A brief history of propaganda during conflict: a lesson for counter-terrorism strategic communications. Int. Center Counter Terror. Hague. doi: 10.19165/2016.1.06

  • CrossRef
  • Google Scholar

• 52

  JääskeläinenP.HuhtinenA. M. 2020. Online polarization, radicalisation, and conspiracy. European conference on information warfare and security, ECCWS

  • Google Scholar

• 53

  JafaY. S. (2005). Defeating terrorism: a study of operational strategy and tactics of police forces in Jammu & Kashmir (India). Police Pract. Res.6, 141–164. doi: 10.1080/15614260500121138

  • CrossRef
  • Google Scholar

• 54

  KarpovaA.SavelevA.VilninA.KuznetsovS. (2022). Method for detecting far-right extremist communities on social media. Soc. Sci.11. doi: 10.3390/socsci11050200

  • CrossRef
  • Google Scholar

• 55

  KhaireD. (2023). Reintegrating vulnerable youth in Jammu and Kashmir: a deradicalization process model by Maharashtra police in India. J. South Asian Stud.11, 171–182. doi: 10.33687/jsas.011.02.4612

  • CrossRef
  • Google Scholar

• 56

  KhosraviNikM.AmerM. (2022). Social media and terrorism discourse: the Islamic state’s (IS) social media discursive content and practices. Crit. Discourse Stud.19, 124–143. doi: 10.1080/17405904.2020.1835684

  • CrossRef
  • Google Scholar

• 57

  KlingbergS. (2022). “Countering terrorism: digital policing of open source intelligence and social updates media using artificial intelligence” in Artificial intelligence and National Security, Amsterdam: IOS Press. 101–111.

  • Google Scholar

• 58

  KlingerU.Lance BennettW.KnüpferC. B.MartiniF.ZhangX. (2023). From the fringes into mainstream politics: intermediary networks and movement-party coordination of a global anti-immigration campaign in Germany. Inf. Commun. Soc.26, 1890–1907. doi: 10.1080/1369118X.2022.2050415

  • CrossRef
  • Google Scholar

• 59

  KornerBottcher (2019). Digital politics AI, big data and the future of democracy for. Perspect. Polit.

  • Google Scholar

• 60

  KumarA.TaylorJ. W. (2024). Feature importance in the age of explainable AI: case study of detecting fake news & misinformation via a multi-modal framework. Eur. J. Oper. Res.317, 401–413. doi: 10.1016/j.ejor.2023.10.003

  • CrossRef
  • Google Scholar

• 61

  LegideK. K. (2022). Exploring the challenges and limits in the compliance with transitional justice norm in non-regime transitions: the case of post-2018 Ethiopia. J. Law Confl. Resolut.13, 1–26. doi: 10.5897/jlcr2022.0338

  • CrossRef
  • Google Scholar

• 62

  LittleR.FordP.GirardiA. (2021). Online self-radicalisation: a case study of cognitive vulnerabilities for radicalisation to extremism and single actor terrorism. J. Intellect. Disabil. Off. Behav.12, 112–123. doi: 10.1108/JIDOB-03-2021-0006

  • CrossRef
  • Google Scholar

• 63

  LombardoS. (2020). “The bad, the good, and the rebellious bots: world’s first in artificial intelligence” in Analyzing future applications of AI, sensors, and robotics in society, Amsterdam: IOS Press. 221–237.

  • Google Scholar

• 64

  MaguR.JoshiK.LuoJ. 2017. Detecting the hate code on social media. Proceedings of the 11th international conference on web and social media, ICWSM 2017.

  • Google Scholar

• 65

  MaskaliūnaitėA. (2015). Exploring the theories of radicalisation. Int. Stud. Interdiscip. Polit. Cult. J.17, 9–26. doi: 10.1515/ipcj-2015-0002

  • CrossRef
  • Google Scholar

• 66

  MatthesJ.KaskeleviciuteR.KnupferH.MasoodM. (2023). The affective nexus between refugees and terrorism: a panel study on how social media use shapes negative attitudes toward refugees. Polit. Psychol.45, 961–978. doi: 10.1111/pops.12950

  • CrossRef
  • Google Scholar

• 67

  McElreathD.DossD. A.McElreathL.LindsleyA.LuskG.SkinnerJ.et al. (2018). The communicating and marketing of radicalism: a case study of isis and cyber recruitment. Int. J. Cyber Warf. Terror.8, 26–45. doi: 10.4018/IJCWT.2018070103

  • CrossRef
  • Google Scholar

• 68

  McKendrickK. (2019). Artificial intelligence prediction and counterterrorism. London: Chatham House.

  • Google Scholar

• 69

  MetaxasP. T. (2020). Technology, propaganda, and the limits of the human intellect. Cambridge, MA: MIT Press. 245–256.

  • Google Scholar

• 70

  MirchandaniM. (2018). Digital hatred, real violence: Majoritarian radicalisation and social media in India (ORF Occasional Paper). New Delhi: Observer Research Foundation.

  • Google Scholar

• 71

  MølmenG. N.RavndalJ. A. (2023). Mechanisms of online radicalisation: how the internet affects the radicalisation of extreme-right lone actor terrorists. Behav. Sci. Terror. Polit. Aggress.15, 463–487. doi: 10.1080/19434472.2021.1993302

  • CrossRef
  • Google Scholar

• 72

  MonaghanJ.MolnarA. (2016). Radicalisation theories, policing practices, and “the future of terrorism?”. Crit. Stud. Terror.9, 393–413. doi: 10.1080/17539153.2016.1178485

  • CrossRef
  • Google Scholar

• 73

  MonarJ. (2015). The EU as an international counter-terrorism actor: progress and constraints. Intell. Natl. Secur.30, 333–356. doi: 10.1080/02684527.2014.988448

  • CrossRef
  • Google Scholar

• 74

  MontasariR.CarrollF.MitchellI.HaraS.Bolton-KingR. (2022). “Privacy, security and forensics in the internet of things (IoT)” in Privacy, security and forensics in the internet of things (IoT). Amsterdam: IOS Press.

  • Google Scholar

• 75

  MozurP.ur-RehmanZ. (2021). How the Taliban turned social media into a tool for control. New York, NY: The New York Times.

  • Google Scholar

• 76

  MughalR.DeMarinisV.NordendahlM.LoneH.PhillipsV.Boyd-MacMillanE. (2023). Public mental health approaches to online radicalisation: an empty systematic review. Int. J. Environ. Res. Public Health20:6586. doi: 10.3390/ijerph20166586,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 77

  NeelamalarM.VivakaranM. V. (2019). A critical analysis of the jihadi discourse through online magazines with special reference to ‘Wyeth’ magazine. India Q.75, 456–471. doi: 10.1177/0974928419874548

  • CrossRef
  • Google Scholar

• 78

  NeoL. S. (2022). Understanding the community’s perceptions towards online radicalisation: an exploratory analysis. Int. J. Cyber Warf. Terror.12, 1–15. doi: 10.4018/IJCWT.297860

  • CrossRef
  • Google Scholar

• 79

  NirupamaB. K. (2023). A systematic literature analysis on the fake identification. Int. J. Sci. Res. Eng. Manag.7. doi: 10.55041/ijsrem25004

  • CrossRef
  • Google Scholar

• 80

  OktavianusJ.DavidsonB. (2023). Countering terrorism on social media: an analysis of online anti-terrorism movement in Indonesia. Commun. Public8, 308–323. doi: 10.1177/20570473231189898

  • CrossRef
  • Google Scholar

• 81

  OlomojobiY.OmotolaO. T. (2021). Social media: a protagonist for terrorism. Int. J. Cyber Warf. Terror.11, 31–44. doi: 10.4018/IJCWT.2021010103

  • CrossRef
  • Google Scholar

• 82

  OmotoyinboF. R. (2014). Online radicalisation: the net or the netizen?Soc. Technol.4, 51–61. doi: 10.13165/st-14-4-1-04

  • CrossRef
  • Google Scholar

• 83

  PandeyR.PandeyM.NazarovA.2022. Detection of propaganda in information warfare using deep learning. Proceedings - 2022 4th international conference on advances in computing, communication control and networking, ICAC3N 2022.

  • Google Scholar

• 84

  PandyaA. (2020). The threat of transnational terrorist groups in Kashmir. Perspect. Terror.14, 13–25.

  • Google Scholar

• 85

  ParvezS.HastingsJ. V. (2022). Understanding the role of digital media in female participation in terrorism: the case of Bangladesh. Small Wars Insurg.33, 1345–1371. doi: 10.1080/09592318.2022.2109371

  • CrossRef
  • Google Scholar

• 86

  PelzerR. (2018). Policing of terrorism using data from social media. Eur. J. Secur. Res.3, 163–179. doi: 10.1007/s41125-018-0029-9

  • CrossRef
  • Google Scholar

• 87

  PerešinA. (2014). Al-Qaeda online radicalisation and the creation of children terrorists. Medij. Istraz.20, 85–101.

  • Google Scholar

• 88

  PfeifferC. P. (2012). Terrorism and its oxygen: a game-theoretic perspective on terrorism and the media. Behav. Sci. Terrorism Polit. Aggress.4, 212–228. doi: 10.1080/19434472.2011.594629

  • CrossRef
  • Google Scholar

• 89

  PiazzaJ. A. (2022). Fake news: the effects of social media disinformation on domestic terrorism. Dyn. Asymmetric Confl.15, 55–77. doi: 10.1080/17467586.2021.1895263

  • CrossRef
  • Google Scholar

• 90

  ProctorD. 2021. Cybernetics and digital whiteness: exposure to radicalisation through feedback loops. 2021 IEEE conference on Norbert wiener in the 21st century: Being human in a Global Village, 21CW 2021.

  • Google Scholar

• 91

  ProrokovićD.ParezanovićM. (2023). Artificial intelligence and psychological: propaganda operations in the context of threat to national security. Politika Nacionalne Bezbednosti25, 13–32. doi: 10.5937/pnb25-46741

  • CrossRef
  • Google Scholar

• 92

  PutraM. D. L. (2017). New media and terrorism: role of the social media to countering cyber terrorism and cyber extremism for effective response. SSRN Electron. J. doi: 10.2139/ssrn.2754370

  • CrossRef
  • Google Scholar

• 93

  RaitanenJ.OksanenA. (2019). Deep interest in school shootings and online radicalisation. J. Threat Assess. Manag.6, 159–172. doi: 10.1037/tam0000127

  • CrossRef
  • Google Scholar

• 94

  RajaW.TantrayT. L.RaniN. (2025). Radicalisation in Jammu and Kashmir: a theoretical and empirical appraisal of internal and external dynamics. Round Table, 1–15. doi: 10.1080/00358533.2025.2561106,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 95

  RajtmajerS.SusserD. 2020. Automated influence and the challenge of cognitive security. ACM international conference proceeding series.

  • Google Scholar

• 96

  RanaM. S.NobiM. N.MuraliB.SungA. H. (2022). Deepfake detection: a systematic literature review. IEEE Access10, 25494–25513. doi: 10.1109/ACCESS.2022.3154404

  • CrossRef
  • Google Scholar

• 97

  RevellT. (2017). How to turn Facebook into a weaponised AI propaganda machine. New Sci.

  • Google Scholar

• 98

  RubinV. L. (2022). “Misinformation and disinformation: detecting fakes with the eye and AI” in Misinformation and disinformation: Detecting fakes with the eye and AI. Amsterdam: IOS Press.

  • Google Scholar

• 99

  RudnerM. (2017). “Electronic jihad”: the internet as Al Qaeda’s catalyst for global terror. Stud. Confl. Terror.40, 10–23. doi: 10.1080/1057610X.2016.1157403

  • CrossRef
  • Google Scholar

• 100

  SaadatS. Y. (2020). International cooperation for counter-terrorism: a strategic perspective. J. Polic. Intell. Counter Terr.15, 83–93. doi: 10.1080/18335330.2020.1732451

  • CrossRef
  • Google Scholar

• 101

  SahelS. (2020). Taliban, terrorism and war on terror: assessing US involvement in Afghanistan. Kardan J. Soc. Sci. Humanit. doi: 10.31841/kjssh.2021.39

  • CrossRef
  • Google Scholar

• 102

  SánchezJ. L. M.RuizM. J. U. (2020). Artificial intelligence and journalism: a tool to fight disinformation. Rev. CIDOB d’Afers Int.124. doi: 10.24241/RCAI.2020.124.1.49

  • CrossRef
  • Google Scholar

• 103

  SaxenaA. K. (2024). Navigating the AI landscape: sectoral insights on integration and impact. Eng. Int.12, 9–28. doi: 10.18034/ei.v12i1.688

  • CrossRef
  • Google Scholar

• 104

  SchnaderJ. (2020). The implementation of artificial intelligence in hard and soft counterterrorism efforts on social media. Santa Clara High Technol. Law J.36.

  • Google Scholar

• 105

  ShabanS. (2020). Teenagers, terrorism, and technopanic: how British newspapers framed female ISIS recruits as victims of social media. Int. J. Commun.14:21.

  • Google Scholar

• 106

  ShahidW.JamshidiB.HakakS.IsahH.KhanW. Z.KhanM. K.et al. (2024). Detecting and mitigating the dissemination of fake news: challenges and future research opportunities. IEEE Trans. Comput. Soc. Syst.11, 4649–4662. doi: 10.1109/TCSS.2022.3177359

  • CrossRef
  • Google Scholar

• 107

  SirgyM. J.EstesR. J.RahtzD. R. (2018). Combatting jihadist terrorism: a quality-of-life perspective. Appl. Res. Qual. Life13, 813–837. doi: 10.1007/s11482-017-9574-z

  • CrossRef
  • Google Scholar

• 108

  SofiW. A.DograA. (2019). Radical space and resistance movement in Maisuma, Srinagar. Asian Rev. Soc. Sci.8, 127–131. doi: 10.51983/arss-2019.8.2.1570

  • CrossRef
  • Google Scholar

• 109

  SpeckhardA.EllenbergM. 2021. Deterrence for Online radicalisation & Recruitment in the 21st century. Amsterdam: IOS Press.

  • Google Scholar

• 110

  SudmannA. (2019). “The democratization of artificial intelligence: net politics in the era of learning algorithms” in The democratization of artificial intelligence: net politics in the era of learning algorithms. Amsterdam: IOS Press.

  • Google Scholar

• 111

  SufiF. (2023). A new social media-driven cyber threat intelligence. Electronics12:1242. doi: 10.3390/electronics12051242

  • CrossRef
  • Google Scholar

• 112

  TabishM.MunirK.AsgharM. F. (2022). Taliban 2.0: an analysis of Taliban’s use of political communication during the Kabul ‘triumph’ and ensuing quest for global recognition. Glob. Foreign Policies Rev.2, 29–38. doi: 10.31703/gfpr.2022(v-ii).04,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 113

  TanoliI.JamilM. L.TanoliI. (2022). Detection of radicalisation and extremism online: a survey. Amsterdam: IOS Press.

  • Google Scholar

• 114

  Thomas-EvansM. (2022). “The internet, social media and the internet of things in radicalisation to terrorism and violent extremism” in Privacy, security and forensics in the internet of things (IoT), Amsterdam: IOS Press. 179–195.

  • Google Scholar

• 115

  TianJ.HuangY.ZhangD. (2022). Detection technology of social robot: based on the interpretation of Botometer model. J. Comput. Sci. Technol. Stud.4, 39–49. doi: 10.32996/jcsts.2022.4.2.6

  • CrossRef
  • Google Scholar

• 116

  TrehanJ. (2002). Terrorism and the funding of terrorism in Kashmir. J. Financ. Crime9, 201–211. doi: 10.1108/eb026018

  • CrossRef
  • Google Scholar

• 117

  TreleavenP.BarnettJ.BrownD.BudA.FenoglioE.KerriganC.et al. (2023). The future of cybercrime: AI and emerging technologies are creating a cybercrime tsunami. SSRN Electron. J. doi: 10.2139/ssrn.4507244

  • CrossRef
  • Google Scholar

• 118

  TulgaA. Y. (2022). Hard and soft terrorism concepts: the case of ISIS. Pak. J. Terror. Res.3, 109–132.

  • Google Scholar

• 119

  WeimannG. (2012). Lone wolves in cyberspace. J. Terror. Res.3. doi: 10.15664/jtr.405

  • CrossRef
  • Google Scholar

• 120

  WilliamsT. J. V.TzaniC. (2024). How does language influence the radicalisation process? A systematic review of research exploring online extremist communication and discussion. Behav. Sci. Terror. Polit. Aggress.16, 310–330. doi: 10.1080/19434472.2022.2104910

  • CrossRef
  • Google Scholar

• 121

  WilliamsonH.FayS.Miles-JohnsonT. (2019). Fear of terrorism: media exposure and subjective fear of attack. Glob. Crime20, 1–25. doi: 10.1080/17440572.2019.1569519

  • CrossRef
  • Google Scholar

• 122

  WilnerA. S.DuboulozC. J. (2011). Transformative radicalisation: applying learning theory to Islamist radicalisation. Stud. Conflict Terrorism34, 418–438. doi: 10.1080/1057610X.2011.561472

  • CrossRef
  • Google Scholar

• 123

  WinterC.NeumannP.Meleagrou-HitchensA.RanstorpM.VidinoL.FürstJ. (2020). Online extremism: research trends in internet activism, radicalisation, and counter-strategies. Int. J. Confl. Violence14. doi: 10.4119/ijcv-3809

  • CrossRef
  • Google Scholar

• 124

  WirsingR. G. (2004). “Introduction: religion, radicalism, and security in South Asia” in Religious radicalism and security in South Asia. Amsterdam: IOS Press.

  • Google Scholar

• 125

  ZennJ. (2023). War on terror 2.0: threat inflation and conflation of far-right and white supremacist terrorism after the capitol “insurrection.”. Crit. Stud. Terror.16, 62–97. doi: 10.1080/17539153.2022.2115218

  • CrossRef
  • Google Scholar

• 126

  ZhaksylykK.BatyrkhanO.ShynarM. 2021. Review of violent extremism detection techniques on social media. Proceedings - 2021 16th international conference on electronics computer and computation, ICECCO 2021.

  • Google Scholar

• 127

  ZvosdetskaO.KuvikS.FedorchukO. (2023). Al-Qaeda’s activities on the internet: forms, types and tools. Mediafor12. doi: 10.31861/mediaforum.2023.12.215-229

  • CrossRef
  • Google Scholar