Boko Haram fighters are increasingly exploiting artificial intelligence tools to improve their operations, using AI-powered chatbots to obtain guidance on battlefield tactics, bomb-making and weapons modification, according to a report by The New York Times.
The report, which cited a study by terrorism and technology researcher Antonia Juelich of the University of Cambridge, said interviews with former Boko Haram members revealed that insurgents had relied on AI to solve operational challenges during attacks in Nigeria.
One former Boko Haram commander recounted that after an assault on a military base was foiled by a defensive trench, the group turned to AI for advice on modifying motorcycles to jump over such obstacles.
According to him, the chatbot generated step-by-step instructions after it was provided with details such as the type of motorcycles used and the distance required for the jump. Mechanics subsequently upgraded the motorcycles for greater speed and acceleration, while fighters rehearsed the manoeuvre before launching another attack.
The study also found that insurgents used AI chatbots to obtain information on designing improvised explosive devices, repairing and upgrading weapons, and gathering operational intelligence.
Former members told the researcher that AI tools provided detailed responses to technical questions, reducing reliance on trial and error.
Another ex-fighter claimed the group also consulted AI to make its explosives more destructive by identifying chemical combinations that increased the impact of bomb blasts.
Some former commanders described AI as an increasingly valuable operational tool, with one reportedly saying, “God has helped us, and so will AI.”
The report said the findings indicate that extremist groups are expanding their use of generative AI beyond propaganda, recruitment and translation into operational planning and battlefield support.
Former insurgents identified several AI platforms they allegedly used, including ChatGPT, Claude, Gemini, Grok and DeepSeek.
The report also cited studies warning that terrorist organisations are experimenting with AI for reconnaissance, communications, coding and intelligence gathering, despite safety measures built into many chatbot systems.
Reacting to the findings, OpenAI said using its products for terrorism or violence violates its policies and that it continues to strengthen safeguards against misuse. Google and Anthropic also stated that their AI models are designed to reject dangerous requests while continuously improving safety protections.
However, the study said experienced insurgents sometimes bypassed chatbot restrictions by disguising harmful requests as legitimate projects, such as claiming the information was intended for film production. Researchers added that the group compared responses across multiple AI platforms to identify systems that disclosed more information.
Although analysts said AI is unlikely to fundamentally transform terrorism in the immediate future, they warned that the technology could enhance the capabilities of lower-level operatives and accelerate the spread of technical knowledge among extremist groups.
Boko Haram insurgents rely on AI to solve operational challenges – Reports
