AI & TERRORISMWill Generative AI Fundamentally Change Terrorist Threats?
Does generative artificial intelligence (Gen AI) represent a genuine transformation in terrorist capability, or are its risks being overstated within contemporary security discourse? Gen AI tools, particularly large language models, can improve the efficiency, accessibility, and scale of certain terrorist activities, but there remains limited evidence that they fundamentally alter the nature of terrorism or significantly enhance operational capability.
Abstract: This article examines whether generative artificial intelligence (Gen AI) represents a genuine transformation in terrorist capability or whether its risks are being overstated within contemporary security discourse. Drawing on affordance theory, it explores how terrorists exploit technologies in ways not intended by their designers and assesses the practical implications of Gen AI across propaganda, recruitment and radicalization, reconnaissance, and attack planning. The article argues that while Gen AI tools, particularly large language models, can improve the efficiency, accessibility, and scale of certain terrorist activities, there remains limited evidence that they fundamentally alter the nature of terrorism or significantly enhance operational capability. Instead, many anticipated threats remain hypothetical and are often shaped by speculative worst-case scenarios rather than demonstrated use. The article further argues that terrorists historically favor technologies that are cheap, accessible, reliable, and easy to adapt, rather than highly sophisticated innovations.
Technology is, for obvious reasons, a major focus for counterterrorism analysts. Modern terrorism originated in the second half of the 19th century partly from the convergence of developments in weaponry, such as the invention of dynamite (1867), and in communication technologies, such as electric telegraphy (patented 1837; first deployed 1844), steam-powered rotary printing presses (1843-47) and Linotype (1866).1 The vast expansion of commercial aviation in the 1960s gave terrorists something new and spectacular to attack, while the arrival of television in people’s homes at around the same time provided them with the means to provide the spectacle.2 Digital technologies from the 1990s enabled terrorist networks to supersede closely knit organizations and achieve global reach; social media and smartphones allowed terrorist brands to exist virtually as well as on the ground.3 As various forms of artificial intelligence are now making real and, in some cases, profound changes to our lives, the public realm is becoming filled with discussion of their potential benefits and risks. It is therefore only to be expected that terrorism analysts have turned toward the potential downsides of these remarkable technologies, and of generative AI (Gen AI) in particular. This article considers the terrorist risk presented by widespread, freely available Gen AI tools, and specifically whether Gen AI offers, as some fear, a step-change in terrorist capability.
