ISIS and Boko Haram use AI chatbots for attack planning

A new study from Cambridge reveals that terrorist groups like ISIS and Boko Haram are actively using major AI chatbots—ChatGPT, Claude, Gemini, Grok, Meta AI, and DeepSeek—for attack planning, weapons development, and operational security. The research, based on interviews with 27 former Boko Haram members, found that ISIS has been offering prompt engineering and jailbreak training since 2023, teaching commanders how to bypass safety filters. Both Boko Haram factions have set up dedicated AI units, with members gathering in rooms to project chatbot usage on big screens. In one bizarre case, they used AI to replicate motorcycle stunts from a movie, resulting in 18 deaths during training. While current use remains conventional, former members expressed strong enthusiasm, and some discussed pursuing AI for mass-casualty weapons. Safety filters are failing to prevent misuse, and researchers warn that the real danger lies not in general chatbots but in specialized AI systems, especially in life sciences. Why it matters: This confirms long-standing fears that current AI safety measures are insufficient against determined adversaries, making the need for robust, enforceable regulation more urgent than ever.