How the terrorist group Boko Haram uses frontier AI
Points and comments are a snapshot, not live.
Report finds Boko Haram systematically uses frontier AI for attack planning and bomb design.
A Cambridge CASP report based on interviews with 27 former Boko Haram members in 2025-2026 finds both factions use ChatGPT, Claude, Gemini, Grok, Meta AI, and DeepSeek for combat and daily operations. Uses include attack planning, weapons troubleshooting, and bomb design, with users circumventing safeguards. Know-how was transferred via transnational jihadist networks, with Islamic State operatives providing in-person training. Respondents expressed enthusiasm for AI and openness to mass-casualty weapons.
Terrorist adoption of AI is more systematic than previously recognized, warranting attention from policymakers and security communities.
What commenters are saying
Commenters are skeptical that AI provides material advantage over existing bomb-making knowledge. They argue jailbroken responses are not actionable and that information is already available on Wikipedia and YouTube. Some note that circumvention tricks like movie script framing no longer work. Others point out that terrorists may be using AI mainly for translation, as LLMs excel at language tasks. A subthread discusses whether terrorists are intelligent or just vulnerable to indoctrination, with examples from cults and Nazi Germany.
A few comments claim that telling an AI you are in Africa can bypass safety measures for legal queries, but this is contested.