A fifth face? The evolving threat of nuclear terrorism in the age of artificial intelligence
Nonproliferation Review, vol.33, no.1-3, pp.65-88, 2026 (Scopus)
- Publication Type: Article / Article
- Volume: 33 Issue: 1-3
- Publication Date: 2026
- Doi Number: 10.1080/10736700.2026.2658838
- Journal Name: Nonproliferation Review
- Journal Indexes: Scopus, Political Science Complete, Political Science Abstract (IPSA)
- Page Numbers: pp.65-88
- Keywords: artificial intelligence, dirty bomb, emerging technologies, nuclear security, nuclear terrorism, nuclear weapons
- Hacettepe University Affiliated: Yes
Abstract
Rapid development of artificial intelligence (AI) presents an underexamined challenge to nuclear facilities. This article analyzes AI’s impact on the threat of nuclear terrorism, asking whether nuclear security simply needs to be updated for the AI age, or whether the arrival of this transformative technology poses a more fundamental shift in the nature of the threat itself. The article approaches this research question by reevaluating the “four faces of nuclear terrorism” typology introduced by Charles D. Ferguson, William C. Potter, and co-authors—widely accepted as orthodoxy in the field of nuclear security. It argues that while AI substantially raises the risk of nuclear terrorism, it does so as a cross-cutting enabler, not as a new conceptual “fifth face.” AI technologies can lower technical and logistical thresholds for nonstate actors seeking nuclear capabilities by enhancing intelligence analysis, automating complex design modeling, and facilitating covert procurement—amplifying preexisting risk vectors. This challenges the position of nuclear “skeptics” that practical difficulties render nuclear terrorism a distant hypothetical. The article also demonstrates how AI may make scenarios like sabotage of nuclear facilities or detonation of a radiological “dirty bomb” more plausible. Ultimately, the AI age necessitates a fundamental reassessment of nuclear terrorism risk frameworks.