Former Boko Haram and Islamic State West Africa Province (ISWAP) commanders reportedly used generative AI chatbots to help plan operations, refine battlefield tactics, and answer technical questions related to explosives.
According to The New York Times, citing interviews by University of Cambridge terrorism and technology researcher Antonia Juelich, former insurgents relied on AI models including ChatGPT, Claude, Gemini, Grok, and DeepSeek to support various stages of military planning.
How AI Entered the Operational Cycle
Juelich’s research draws on nearly 60 interviews with 27 former Boko Haram members conducted in Nigeria over the past year.
The findings suggest the groups integrated generative AI across the operational cycle, from mission planning and tactical decision-making to post-mission analysis.
The reported use extends beyond the propaganda, recruitment, and translation tasks that have typically defined discussions of extremist AI adoption.
Interviewees described using AI to solve practical battlefield problems.
One former Boko Haram commander said the group consulted an AI chatbot after a planned assault on a military base was thwarted by a defensive trench. Inspired by a movie scene, members reportedly asked the chatbot how to modify motorcycles to cross the obstacle before testing the changes during training.
Former ISWAP commanders also said they used AI to answer technical questions about explosives, claiming the chatbots reduced reliance on trial and error.
Circumventing AI Safeguards
According to the report, the former insurgents often bypassed chatbot safeguards by framing prompts as academic or engineering questions rather than requests related to violence.
OpenAI said using its models to facilitate terrorism or violence violates its policies, while other AI developers told The New York Times that their systems are designed to reject dangerous requests.
Juelich said terrorist groups’ use of generative AI has been “significantly underestimated in both scope and character,” highlighting growing concerns that commercially available AI tools are becoming part of operational planning.
