AI-enabled terrorism refers to the adoption of AI systems — in the documented cases, general-purpose frontier chatbots — by terrorist and extremist organizations as operational tools: designing explosives, repairing and upgrading weapons, planning attacks, and producing propaganda. Counterterrorism literature has organized around the misuse of AI by violent actors since roughly 2018, but documented, institutional-level adoption by a named terrorist organization was first reported in detail in July 2026. The category is distinct from AI-driven political violence, in which AI development is the target of violence rather than a tool of it.
Documented cases
Boko Haram (2026)
A New York Times investigation published July 10, 2026 reported that members of the extremist group Boko Haram are using AI chatbots to design explosives, repair and upgrade weapons, and brainstorm attack ideas (Source: nytimes.com).
A report by the Cambridge Programme on AI Science & Policy (CASP), an interdisciplinary research programme on frontier AI at the University of Cambridge, circulated July 13, 2026 — "'God has helped us, and so will AI': How the Terrorist Group Boko Haram Uses Frontier AI," by Antonia Juelich — found, based on semi-structured interviews with 27 former Boko Haram members conducted in northeast Nigeria in 2025 and 2026, that both of the group's factions institutionally use ChatGPT, Claude, Gemini, Grok, Meta AI, and DeepSeek for attack planning and explosives troubleshooting, with training in the tools transferred via Islamic State operatives ('God has helped us, and so will AI': How the Terrorist Group Boko Haram Uses Frontier AI (CASP, 2026)). The report's abstract describes AI-assisted activity primarily through 2024, institutionalized through specialized units and internal training, with users successfully circumventing some safeguards; it notes that respondents expressed enthusiasm for AI and in some cases openness to mass-casualty weapons, though documented use remains conventional ('God has helped us, and so will AI': How the Terrorist Group Boko Haram Uses Frontier AI (CASP, 2026)). An earlier version of this page expanded CASP as "Centre for the Study of Armed Groups and Proliferation"; the programme's own site gives the name as the Cambridge Programme on AI Science & Policy.
Mechanisms and policy relevance
The reported uses cluster around uplift to existing operations — troubleshooting explosives, maintaining captured weapons, and generating plans — rather than novel weapons categories. The pattern bears on several existing policy frames:
- Safeguard evasion. The reported use spans models from six developers, closed- and open-weight, indicating that refusal training and safety classifiers on any single model do not remove access at the category level. See Jailbreaking and Red Teaming and Open-Weight Frontier Models.
- Threat-actor assumptions. Frontier-lab safety frameworks have reasoned that terrorist groups remain bottlenecked by technical skill and resources (see CBRN Uplift and the GPT-5.6 preview system card's reuse of the GPT-5.3 Codex threat model on GPT-5.6 (Sol, Terra, Luna)); the CASP finding of institutionalized chatbot use with skills transfer between organizations is direct evidence on how that bottleneck is shifting for conventional (non-CBRN) capability.
- Dual-use governance. The cases sit within the broader Dual-Use Frontier AI debate over whether generally available models materially uplift violent non-state actors.
Relationships
- related: AI-Driven Political Violence — the inverse category (AI as target of violence, not tool)
- related: Dual-Use Frontier AI, CBRN Uplift, Jailbreaking and Red Teaming, AI and National Security
- related: Autonomous cyber-agents — parallel misuse thread in the cyber domain