September 2024
In its 213th book, “Terrorism and Artificial Intelligence: Tools, Confrontations, and the Future,” published September 2024, the Al-Mesbar Studies and Research Center explores how terrorist organizations use artificial intelligence (AI) in their ideological campaigns and recruitment efforts. The book identifies methods of AI usage by terrorist groups and examines the current and future risks this poses to national security and social stability, especially in a world where digital progress is increasingly central. It also suggests a number of strategies by which government bodies and civil society can counter these threats.
The first study, by Egyptian researcher Mahmoud El-Tabbakh, investigates the dual use of generative AI by terrorists. This technology amplifies their influence by enhancing traditional techniques of encryption and concealment. Analyses by the “Tech against Terrorism” initiative show that extremist groups have used AI tools to produce over 5,000 distinct propaganda materials, including images, cartoons, and inciting songs. AI also facilitates more sophisticated cyber-attacks, such as ISIS’s “Caliphate Cannon” campaign, which launched Distributed Denial of Service (DDoS) attacks targeting critical infrastructure. Advances in machine learning allow automation of these attacks, increasing cyber threat levels.
Simultaneously, terrorist groups are seeking to exploit autonomous vehicles as potential weapons, building on the success of the traditional car bombs, as for example those used in the 2016 Berlin and 2017 Barcelona attacks. This development opens the possibility for more advanced scenarios involving smart vehicles.
Terrorists also create fake news platforms. For example, ISIS’s Khorasan Province runs a “deepfake” disinformation campaign, while ISIS central employs AI to generate jihadist chants to recruit new followers. These materials appear on apps popular with teenagers and can generate high-quality images and animations instantly, without professional designers or lengthy production timelines.
Recruitment methods have also grown increasingly sophisticated. Groups linked to Al-Qaeda published an e-book in February 2024 titled “Amazing Ways to Use AI-Based Chatbots,” serving as a practical guide to exploiting AI for influence and recruitment. Extremists hacked Meta’s LLaMa model to develop chatbots specialized in propagating extremist ideology. AI personas like “Trinity” have emerged, promoting hate speech and racial discrimination in myriad ways.
Today we are witnessing an intensifying competition between terrorist groups and security agencies in AI use. Extremist organizations rapidly adopt new technologies and adapt to countermeasures. The main challenge is anticipating digital terrorism strategies and preventing exploitation of modern technologies. Traditional solutions are insufficient; stakeholders must enhance AI expertise and develop advanced tools to detect and prevent extremist activities before they cause harm.
Researchers at the American Counterterrorism Targeting and Resilience Institute—Arian Shajkoi, Michael Vandelune, and Allison McDowell-Smith—identify methods and virtual private networks which ISIS and other radical groups use to evade surveillance and conceal online activities. They note increased use of AI and advanced techniques in propaganda, operational planning, and recruitment. These groups develop targeted strategies, analyze large data in real time, and automate tasks.
Stuart Macdonald, Professor of Law and Counterterrorism at Swansea University, focuses on AI and monitoring terrorist content online. He reviews various automated content monitoring tools, especially those using AI. His study outlines two main approaches: matching-based detection and classification-based detection, before discussing behavior-based detection. These tools often combine in a multi-layered approach. Macdonald emphasizes that human involvement is essential since matching-based methods require continuously identifying terrorist content for a fingerprint database, and classification methods need large, curated, and labeled datasets to train machine learning algorithms.
Most classification tools flag content for human review because humans better assess context, subtle differences, intent, and social, cultural, historical, and political factors. Humans also handle appeals when automated tools produce false positives. Macdonald concludes that AI is necessary but, on its own, insufficient to combat terrorist content online.
Egyptian researcher Salma Saqr examines how terrorist groups and far-right extremists use generative AI and how to counter these threats. She highlights unconventional security risks linked to rapid technological advances, especially in finance. Terrorists digitize donations and fundraising, raising amounts far beyond traditional methods. Extremist influencers use various online platforms to solicit purchases and donations, with cryptocurrencies serving as ideal payment systems, alongside e-commerce platforms.
Researcher and academic Hakim Gharib identifies key AI technologies and roles in reducing terrorist actions, including combating extremist ideas, predicting attacks, identifying terrorists’ faces, brain fingerprinting (brainwave scanning), analytical sensing surveillance, and crowd analysis.
The book also reveals how Islamist groups exploit religious texts via digital applications to reinforce terrorist ideology. Mustafa Qutb investigates the use of algorithms in Islamic sciences, focusing on natural language processing in Islamic texts, machine learning as applied to fatwas, algorithmic interpretation, machine translation, and controls for applying algorithms to Islamic studies.
Researchers Leandra Robert and Emad Naim study AI’s role in art and literature and how to use AI techniques to counter terrorism and counter violent extremism. For instance, the Deep Dream program transforms images into surreal forms, revolutionizing animation. Refik Anadol uses big data in digital sculpture, and Sofia Crespo developed the Neural Zoo project, generating hybrid creatures from natural data.
In the Arab world, projects like “Araby.ai” merge digital art with Arab culture, and several programs have been developed to revive Arabic calligraphy.
AI’s influence on literature began with a Japanese team that developed the novel The Day the Computer Writes a Novel in 2016, which reached the finals of a literary competition.
In the Arab context, Mohamed Ahmad Fouad wrote the novel “The Last Lives of Beings” with AI assistance in under 23 hours, sparking debate about the future of writing.
In cinema, the screenplay for the 2016 movie “Sunspring” was written entirely by AI. However, to date, the authors conclude that AI remains a tool supporting human creativity, not a replacement for the human element in shaping cinematic vision.
AI raises legal challenges. A U.S. court refused to grant copyright protection for artwork produced solely by AI, stating that human creativity is required for legal protection. Questions remain about intellectual property ownership—whether it belongs to the programmer, the user, or the developer company.
Egyptian researcher Alaa Al-Hamd studies how literature created with AI techniques can combat terrorism. Terrorist groups often use novels and literature online to spread extremist ideas, whereas mainstream novelists can counter this by including anti-terrorist themes in their works.
Academic and researcher Ahmed Abdel Majeed explores journalism’s future amid AI challenges. He highlights expert opinions in computer science and digital technology on the future of journalism, ranging from optimistic to pessimistic. He presents three future scenarios for AI journalism: creativity, stability, and pessimism.
In conclusion, Al Mesbar Studies and Research Center thanks the contributing researchers and those involved in publishing this book, hoping it fills a gap in the Arab library.
Editor-in-Chief
Omar Al-Bashir Al-Turabi
September 2024
