A new report has sparked global concern over the potential misuse of artificial intelligence, alleging that Boko Haram and its splinter faction, the Islamic State West Africa Province (ISWAP), used AI-powered chatbots to improve aspects of their operations.
According to The New York Times, former insurgent commanders told terrorism researcher Antonia Juelich of the University of Cambridge that members of the groups consulted AI platforms, including ChatGPT, Claude, Gemini, Grok and DeepSeek, while looking for ideas to overcome battlefield challenges.
The report states that after an assault on a military base was slowed by a defensive trench, the group reportedly used AI tools to explore ways of modifying motorcycles to cross the obstacle. Former commanders said mechanics altered the bikes and fighters later practised the manoeuvres before operations.
The former insurgents also claimed they were sometimes able to bypass chatbot safeguards by disguising dangerous requests as educational or research-related questions. These claims have renewed concerns about how emerging AI technologies could be exploited by violent extremist groups.
In response, OpenAI said using its services to support terrorism violates its policies. Google and Anthropic also said their AI systems include safeguards designed to reject requests that could facilitate harmful or dangerous activities.
Security experts say the report highlights the growing challenge of ensuring advanced AI technologies cannot be exploited for violent purposes, even as developers continue strengthening safety protections.
