Boko Haram uses AI chatbots for bomb making and attack planning

A University of Cambridge study led by terrorism researcher Antonia Juelich found that Boko Haram insurgents in Nigeria have turned to frontier artificial‑intelligence tools—including ChatGPT, Claude, Gemini, Grok and DeepSeek—to obtain technical guidance for building improvised explosive devices, upgrading weapons and devising battlefield tactics. 57 interviews with 27 former fighters conducted between 2025 and 2026 revealed that commanders used AI to learn how to modify motorcycles to jump defensive trenches, to increase bomb potency by selecting more powerful chemical mixes, and to receive step‑by‑step instructions for weapon fabrication. The group reportedly bypassed chatbot safety guardrails by “jailbreaking” prompts, disguising violent requests as legitimate projects. The findings were reported by The New York Times and other outlets. OpenAI, Google and Anthropic said their models contain protections against such misuse, but researchers noted that insurgents were able to obtain useful information in many cases. While the study lacks independent forensic evidence, it documents how extremist groups are experimenting with generative AI beyond propaganda, potentially enhancing the capabilities of lower‑level operatives in the ongoing Boko Haram insurgency, which has already caused more than 35,000 deaths and displaced over two million people.